Introduction to Service Telemetry Framework 1.4

Service Telemetry Framework (STF) collects monitoring data from OpenStack (OSP) or third-party nodes. You can use STF to perform the following tasks:

  • Store or archive the monitoring data for historical information.

  • View the monitoring data graphically on the dashboard.

  • Use the monitoring data to trigger alerts or warnings.

The monitoring data can be either metric or event:

Metric

A numeric measurement of an application or system.

Event

Irregular and discrete occurrences that happen in a system.

The components of STF use a message bus for data transport. Other modular components that receive and store data are deployed as containers on OpenShift.

Additional resources

Service Telemetry Framework architecture

Service Telemetry Framework (STF) uses a client-server architecture, in which OpenStack (OSP) is the client and OpenShift is the server.

STF consists of the following components:

  • Data collection

    • collectd: Collects infrastructure metrics and events.

    • Ceilometer: Collects OSP metrics and events.

  • Transport

    • Apache Qpid Dispatch Router: An AMQP 1.x compatible messaging bus that provides fast and reliable data transport to transfer the metrics to STF for storage.

    • Smart Gateway: A Golang application that takes metrics and events from the AMQP 1.x bus to deliver to ElasticSearch or Prometheus.

  • Data storage

    • Prometheus: Time-series data storage that stores STF metrics received from the Smart Gateway.

    • ElasticSearch: Events data storage that stores STF events received from the Smart Gateway.

  • Observation

    • Alertmanager: An alerting tool that uses Prometheus alert rules to manage alerts.

    • Grafana: A visualization and analytics application that you can use to query, visualize, and explore data.

The following table describes the application of the client and server components:

Table 1. Client and server components of STF
Component Client Server

An AMQP 1.x compatible messaging bus

yes

yes

Smart Gateway

no

yes

Prometheus

no

yes

ElasticSearch

no

yes

collectd

yes

no

Ceilometer

yes

no

To ensure that the monitoring platform can report operational problems with your cloud, do not install STF on the same infrastructure that you are monitoring.
Service Telemetry Framework architecture overview
Figure 1. Service Telemetry Framework architecture overview

For client side metrics, collectd provides infrastructure metrics without project data, and Ceilometer provides OSP platform data based on projects or user workload. Both Ceilometer and collectd deliver data to Prometheus by using the Apache Qpid Dispatch Router transport, delivering the data through the message bus. On the server side, a Golang application called the Smart Gateway takes the data stream from the bus and exposes it as a local scrape endpoint for Prometheus.

If you plan to collect and store events, collectd and Ceilometer deliver event data to the server side by using the Apache Qpid Dispatch Router transport. Another Smart Gateway writes the data to the ElasticSearch datastore.

Server-side STF monitoring infrastructure consists of the following layers:

  • Service Telemetry Framework 1.4

  • OpenShift 4.7 through 4.8

  • Infrastructure platform

Server-side STF monitoring infrastructure
Figure 2. Server-side STF monitoring infrastructure

Installation size of OpenShift

The size of your OpenShift installation depends on the following factors:

  • The infrastructure that you select.

  • The number of nodes that you want to monitor.

  • The number of metrics that you want to collect.

  • The resolution of metrics.

  • The length of time that you want to store the data.

Installation of Service Telemetry Framework (STF) depends on an existing OpenShift environment.

For more information about minimum resources requirements when you install OpenShift on baremetal, see Minimum resource requirements in the Installing a cluster on bare metal guide. For installation requirements of the various public and private cloud platforms that you can install, see the corresponding installation documentation for your cloud platform of choice.

Development environment resource requirements

You can create an all-in-one development environment for STF locally by using CodeReady Containers. The installation process of CodeReady Containers (CRC) is available at https://code-ready.github.io/crc/#installation_gsg.

The minimum resource requirements for CRC is not enough by default to run STF. Ensure that your host system has the following resources available:

  • 4 physical cores (8 hyperthreaded cores)

  • 64 GB of memory

  • 80 GB of storage space

After you complete the installation of CRC, use the crc start command to start your environment. The recommended minimum system resources for running STF in CodeReady Containers is 48 GB of memory and 8 virtual CPU cores:

crc start --memory=49152 --cpus=8

If you have an existing environment, delete it, and recreate it to ensure that the resource requests have an effect.

Procedure
  1. Enter the crc delete command.

    crc delete
  2. Run the crc start command to create your environment:

    crc start --memory=49152 --cpus=8

Preparing your OpenShift environment for Service Telemetry Framework

To prepare your OpenShift environment for Service Telemetry Framework (STF), you must plan for persistent storage, adequate resources, and event storage:

  • Ensure that persistent storage is available in your OpenShift cluster for a production grade deployment. For more information, see Persistent volumes.

  • Ensure that enough resources are available to run the Operators and the application containers. For more information, see Resource allocation.

  • STF uses ElasticSearch to store events, which requires a larger than normal vm.max_map_count. The vm.max_map_count value is set by default in OpenShift. For more information about how to edit the value of vm.max_map_count, see Node tuning operator.

Observability Strategy in Service Telemetry Framework

Service Telemetry Framework (STF) does not include storage backends and alerting tools. STF uses community operators to deploy Prometheus, Alertmanager, Grafana, and Elasticsearch. STF makes requests to these community operators to create instances of each application configured to work with STF.

Instead of having Service Telemetry Operator create custom resource requests, you can use your own deployments of these applications or other compatible applications, and scrape the metrics Smart Gateways for delivery to your own Prometheus-compatible system for telemetry storage. If you set the observability strategy to use alternative backends instead, persistent or ephemeral storage is not required for STF.

Persistent volumes

Service Telemetry Framework (STF) uses persistent storage in OpenShift to request persistent volumes so that Prometheus and ElasticSearch can store metrics and events.

When you enable persistent storage through the Service Telemetry Operator, the Persistent Volume Claims (PVC) requested in an STF deployment results in an access mode of RWO (ReadWriteOnce). If your environment contains pre-provisioned persistent volumes, ensure that volumes of RWO are available in the OpenShift default configured storageClass.

Additional resources

Ephemeral storage

You can use ephemeral storage to run Service Telemetry Framework (STF) without persistently storing data in your OpenShift cluster.

If you use ephemeral storage, you might experience data loss if a pod is restarted, updated, or rescheduled onto another node. Use ephemeral storage only for development or testing, and not production environments.

Resource allocation

To enable the scheduling of pods within the OpenShift infrastructure, you need resources for the components that are running. If you do not allocate enough resources, pods remain in a Pending state because they cannot be scheduled.

The amount of resources that you require to run Service Telemetry Framework (STF) depends on your environment and the number of nodes and clouds that you want to monitor.

Additional resources

Node tuning operator

STF uses ElasticSearch to store events, which requires a larger than normal vm.max_map_count. The vm.max_map_count value is set by default in OpenShift.

If your host platform is a typical OpenShift 4 environment, do not make any adjustments. The default node tuning operator is configured to account for ElasticSearch workloads.

If you want to edit the value of vm.max_map_count, you cannot apply node tuning manually using the sysctl command because OpenShift manages nodes directly. To configure values and apply them to the infrastructure, you must use the node tuning operator. For more information, see Using the Node Tuning Operator.

In an OKD deployment, the default node tuning operator specification provides the required profiles for ElasticSearch workloads or pods scheduled on nodes. To view the default cluster node tuning specification, run the following command:

$ oc get Tuned/default -o yaml -n openshift-cluster-node-tuning-operator

The output of the default specification is documented at Default profiles set on a cluster. You can manage the assignment of profiles in the recommend section where profiles are applied to a node when certain conditions are met. When scheduling ElasticSearch to a node in STF, one of the following profiles is applied:

  • openshift-control-plane-es

  • openshift-node-es

When scheduling an ElasticSearch pod, there must be a label present that matches tuned.openshift.io/elasticsearch. If the label is present, one of the two profiles is assigned to the pod. No action is required by the administrator if you use the recommended Operator for ElasticSearch. If you use a custom-deployed ElasticSearch with STF, ensure that you add the tuned.openshift.io/elasticsearch label to all scheduled pods.

Additional resources

Installing the core components of Service Telemetry Framework

You can use Operators to load the Service Telemetry Framework (STF) components and objects. Operators manage each of the following STF core and community components:

  • Apache Qpid Dispatch Router

  • Smart Gateway

  • Prometheus and AlertManager

  • ElasticSearch

  • Grafana

Prerequisites
  • An OpenShift version inclusive of 4.7 through 4.8 is running.

  • You have prepared your OpenShift environment and ensured that there is persistent storage and enough resources to run the STF components on top of the OpenShift environment. For more information, see Service Telemetry Framework Performance and Scaling.

Additional resources

Deploying Service Telemetry Framework to the OpenShift environment

Deploy Service Telemetry Framework (STF) to collect, store, and monitor events:

Procedure
  1. Create a namespace to contain the STF components, for example, service-telemetry:

    $ oc new-project service-telemetry
  2. Create an OperatorGroup in the namespace so that you can schedule the Operator pods:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1
    kind: OperatorGroup
    metadata:
      name: service-telemetry-operator-group
      namespace: service-telemetry
    spec:
      targetNamespaces:
      - service-telemetry
    EOF

    For more information, see OperatorGroups.

  3. Before you deploy STF on OpenShift, you must enable the catalog source. Install a CatalogSource that contains the Service Telemetry Operator and the Smart Gateway Operator:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: CatalogSource
    metadata:
      name: infrawatch-operators
      namespace: openshift-marketplace
    spec:
      displayName: InfraWatch Operators
      image: quay.io/infrawatch-operators/infrawatch-catalog:nightly
      publisher: InfraWatch
      sourceType: grpc
      updateStrategy:
        registryPoll:
          interval: 30m
    EOF
  4. Validate the creation of your CatalogSource:

    $ oc get -nopenshift-marketplace catalogsource infrawatch-operators
    
    NAME                   DISPLAY                TYPE   PUBLISHER    AGE
    infrawatch-operators   InfraWatch Operators   grpc   InfraWatch   2m16s
  5. Validate that the Operators are available from the catalog:

    $ oc get packagemanifests | grep InfraWatch
    
    service-telemetry-operator                    InfraWatch Operators       7m20s
    smart-gateway-operator                        InfraWatch Operators       7m20s
  6. Enable the OperatorHub.io Community Catalog Source to install data storage and visualization Operators:

    Red Hat supports the core Operators and workloads, including Apache Qpid Dispatch Router, AMQ Certificate Manager, Service Telemetry Operator, and Smart Gateway Operator. Red Hat does not support the community Operators or workload components, inclusive of ElasticSearch, Prometheus, Alertmanager, Grafana, and their Operators.
    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: CatalogSource
    metadata:
      name: operatorhubio-operators
      namespace: openshift-marketplace
    spec:
      sourceType: grpc
      image: quay.io/operatorhubio/catalog:latest
      displayName: OperatorHub.io Operators
      publisher: OperatorHub.io
    EOF
  7. Subscribe to the AMQ Certificate Manager Operator by using the redhat-operators CatalogSource:

    The AMQ Certificate Manager deploys to the openshift-operators namespace and is then available to all namespaces across the cluster. As a result, on clusters with a large number of namespaces, it can take several minutes for the Operator to be available in the service-telemetry namespace. The AMQ Certificate Manager Operator is not compatible with the dependency management of Operator Lifecycle Manager when you use it with other namespace-scoped operators.
    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: amq7-cert-manager-operator
      namespace: openshift-operators
    spec:
      channel: 1.x
      installPlanApproval: Automatic
      name: amq7-cert-manager-operator
      source: redhat-operators
      sourceNamespace: openshift-marketplace
    EOF
  8. Validate your ClusterServiceVersion. Ensure that amq7-cert-manager.v1.0.3 displays a phase of Succeeded:

    $ oc get csv --namespace openshift-operators --selector operators.coreos.com/amq7-cert-manager-operator.openshift-operators
    
    NAME                       DISPLAY                                         VERSION   REPLACES                   PHASE
    amq7-cert-manager.v1.0.3   Red Hat Integration - AMQ Certificate Manager   1.0.3     amq7-cert-manager.v1.0.2   Succeeded
  9. Subscribe to the AMQ Interconnect Operator by using the redhat-operators CatalogSource:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: amq7-interconnect-operator
      namespace: service-telemetry
    spec:
      channel: 1.10.x
      installPlanApproval: Automatic
      name: amq7-interconnect-operator
      source: redhat-operators
      sourceNamespace: openshift-marketplace
    EOF
  10. Validate your ClusterServiceVersion. Ensure that amq7-interconnect-operator.v1.10.4 displays a phase of Succeeded:

    $ oc get csv --selector=operators.coreos.com/amq7-interconnect-operator.service-telemetry
    
    NAME                                 DISPLAY                                  VERSION   REPLACES                             PHASE
    amq7-interconnect-operator.v1.10.4   Red Hat Integration - AMQ Interconnect   1.10.4    amq7-interconnect-operator.v1.10.3   Succeeded
  11. If you plan to store metrics in Prometheus, you must enable the Prometheus Operator. To enable the Prometheus Operator, create the following manifest in your OpenShift environment:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: prometheus
      namespace: service-telemetry
    spec:
      channel: beta
      installPlanApproval: Automatic
      name: prometheus
      source: operatorhubio-operators
      sourceNamespace: openshift-marketplace
    EOF
  12. Verify that the ClusterServiceVersion for Prometheus Succeeded:

    $ oc get csv --selector=operators.coreos.com/prometheus.service-telemetry
    
    NAME                        DISPLAY               VERSION   REPLACES                    PHASE
    prometheusoperator.0.47.0   Prometheus Operator   0.47.0    prometheusoperator.0.37.0   Succeeded
  13. If you plan to store events in ElasticSearch, you must enable the Elastic Cloud on Kubernetes (ECK) Operator. To enable the ECK Operator, create the following manifest in your OpenShift environment:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: elasticsearch-eck-operator-certified
      namespace: service-telemetry
    spec:
      channel: stable
      installPlanApproval: Automatic
      name: elasticsearch-eck-operator-certified
      source: certified-operators
      sourceNamespace: openshift-marketplace
    EOF
  14. Verify that the ClusterServiceVersion for Elastic Cloud on Kubernetes Succeeded:

    $ oc get csv --selector=operators.coreos.com/elasticsearch-eck-operator-certified.service-telemetry
    
    NAME                                         DISPLAY                        VERSION   REPLACES   PHASE
    elasticsearch-eck-operator-certified.1.9.1   Elasticsearch (ECK) Operator   1.9.1                Succeeded
  15. Create the Service Telemetry Operator subscription to manage the STF instances:

    $ oc create -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: service-telemetry-operator
      namespace: service-telemetry
    spec:
      channel: stable-1.4
      installPlanApproval: Automatic
      name: service-telemetry-operator
      source: redhat-operators
      sourceNamespace: openshift-marketplace
    EOF
  16. Validate the Service Telemetry Operator and the dependent operators:

    $ oc get csv --namespace service-telemetry
    
    NAME                                         DISPLAY                                         VERSION        REPLACES                             PHASE
    amq7-cert-manager.v1.0.3                     Red Hat Integration - AMQ Certificate Manager   1.0.3          amq7-cert-manager.v1.0.2             Succeeded
    amq7-interconnect-operator.v1.10.4           Red Hat Integration - AMQ Interconnect          1.10.4         amq7-interconnect-operator.v1.10.3   Succeeded
    elasticsearch-eck-operator-certified.1.9.1   Elasticsearch (ECK) Operator                    1.9.1                                               Succeeded
    prometheusoperator.0.47.0                    Prometheus Operator                             0.47.0         prometheusoperator.0.37.0            Succeeded
    service-telemetry-operator.v1.4.1641489191   Service Telemetry Operator                      1.4.1641489191                                      Succeeded
    smart-gateway-operator.v4.0.1641489202       Smart Gateway Operator                          4.0.1641489202                                      Succeeded

Creating a ServiceTelemetry object in OpenShift

Create a ServiceTelemetry object in OpenShift to result in the Service Telemetry Operator creating the supporting components for a Service Telemetry Framework (STF) deployment. For more information, see Primary parameters of the ServiceTelemetry object.

Procedure
  1. To create a ServiceTelemetry object that results in an STF deployment that uses the default values, create a ServiceTelemetry object with an empty spec parameter:

    $ oc apply -f - <<EOF
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec: {}
    EOF

    To override a default value, define the parameter that you want to override. In this example, enable ElasticSearch by setting enabled to true:

    $ oc apply -f - <<EOF
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        events:
          elasticsearch:
            enabled: true
    EOF

    Creating a ServiceTelemetry object with an empty spec parameter results in an STF deployment with the following default settings:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      alerting:
        alertmanager:
          receivers:
            snmpTraps:
              enabled: false
              target: 192.168.24.254
          storage:
            persistent:
              pvcStorageRequest: 20G
            strategy: persistent
        enabled: true
      backends:
        events:
          elasticsearch:
            enabled: false
            storage:
              persistent:
                pvcStorageRequest: 20Gi
              strategy: persistent
            version: 7.16.1
        logs:
          loki:
            enabled: false
            flavor: 1x.extra-small
            replicationFactor: 1
            storage:
              objectStorageSecret: test
              storageClass: standard
        metrics:
          prometheus:
            enabled: true
            scrapeInterval: 10s
            storage:
              persistent:
                pvcStorageRequest: 20G
              retention: 24h
              strategy: persistent
      clouds:
      - events:
          collectors:
          - collectorType: collectd
            debugEnabled: false
            subscriptionAddress: collectd/cloud1-notify
          - collectorType: ceilometer
            debugEnabled: false
            subscriptionAddress: anycast/ceilometer/cloud1-event.sample
        metrics:
          collectors:
          - collectorType: collectd
            debugEnabled: false
            subscriptionAddress: collectd/cloud1-telemetry
          - collectorType: ceilometer
            debugEnabled: false
            subscriptionAddress: anycast/ceilometer/cloud1-metering.sample
          - collectorType: sensubility
            debugEnabled: false
            subscriptionAddress: sensubility/cloud1-telemetry
        name: cloud1
      graphing:
        enabled: false
        grafana:
          adminPassword: secret
          adminUser: root
          baseImage: docker.io/grafana/grafana:latest
          disableSignoutMenu: false
          ingressEnabled: false
      highAvailability:
        enabled: false
      observabilityStrategy: use_community
      transports:
        qdr:
          enabled: true
          web:
            enabled: false

    To override these defaults, add the configuration to the spec parameter.

  2. View the STF deployment logs in the Service Telemetry Operator:

    $ oc logs --selector name=service-telemetry-operator
    
    ...
    --------------------------- Ansible Task Status Event StdOut  -----------------
    
    PLAY RECAP *********************************************************************
    localhost                  : ok=57   changed=0    unreachable=0    failed=0    skipped=20   rescued=0    ignored=0
Verification
  • To determine that all workloads are operating correctly, view the pods and the status of each pod.

    If you set the backends.events.elasticsearch.enabled parameter to true, the notification Smart Gateways report Error and CrashLoopBackOff error messages for a period of time before ElasticSearch starts.
    $ oc get pods
    
    NAME                                                      READY   STATUS    RESTARTS   AGE
    alertmanager-default-0                                    2/2     Running   0          17m
    default-cloud1-ceil-meter-smartgateway-6484b98b68-vd48z   2/2     Running   0          17m
    default-cloud1-coll-meter-smartgateway-799f687658-4gxpn   2/2     Running   0          17m
    default-cloud1-sens-meter-smartgateway-c7f4f7fc8-c57b4    2/2     Running   0          17m
    default-interconnect-54658f5d4-pzrpt                      1/1     Running   0          17m
    elastic-operator-66b7bc49c4-sxkc2                         1/1     Running   0          52m
    interconnect-operator-69df6b9cb6-7hhp9                    1/1     Running   0          50m
    prometheus-default-0                                      2/2     Running   1          17m
    prometheus-operator-6458b74d86-wbdqp                      1/1     Running   0          51m
    service-telemetry-operator-864646787c-hd9pm               1/1     Running   0          51m
    smart-gateway-operator-79778cf548-mz5z7                   1/1     Running   0          51m

Primary parameters of the ServiceTelemetry object

The ServiceTelemetry object comprises the following primary configuration parameters:

  • alerting

  • backends

  • clouds

  • graphing

  • highAvailability

  • transports

You can configure each of these configuration parameters to provide different features in an STF deployment.

Support for servicetelemetry.infra.watch/v1alpha1 was removed from STF 1.3.

The backends parameter

Use the backends parameter to control which storage back ends are available for storage of metrics and events, and to control the enablement of Smart Gateways that the clouds parameter defines. For more information, see The clouds parameter.

Currently, you can use Prometheus as the metrics storage back end and ElasticSearch as the events storage back end.

Enabling Prometheus as a storage back end for metrics

To enable Prometheus as a storage back end for metrics, you must configure the ServiceTelemetry object.

Procedure
  • Configure the ServiceTelemetry object:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        metrics:
          prometheus:
            enabled: true
Configuring persistent storage for Prometheus

Use the additional parameters that are defined in backends.metrics.prometheus.storage.persistent to configure persistent storage options for Prometheus, such as storage class and volume size.

Use storageClass to define the back end storage class. If you do not set this parameter, the Service Telemetry Operator uses the default storage class for the OpenShift cluster.

Use the pvcStorageRequest parameter to define the minimum required volume size to satisfy the storage request. If volumes are statically defined, it is possible that a volume size larger than requested is used. By default, Service Telemetry Operator requests a volume size of 20G (20 Gigabytes).

Procedure
  • List the available storage classes:

    $ oc get storageclasses
    NAME                 PROVISIONER                RECLAIMPOLICY   VOLUMEBINDINGMODE      ALLOWVOLUMEEXPANSION   AGE
    csi-manila-ceph      manila.csi.openstack.org   Delete          Immediate              false                  20h
    standard (default)   kubernetes.io/cinder       Delete          WaitForFirstConsumer   true                   20h
    standard-csi         cinder.csi.openstack.org   Delete          WaitForFirstConsumer   true                   20h
  • Configure the ServiceTelemetry object:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        metrics:
          prometheus:
            enabled: true
            storage:
              strategy: persistent
              persistent:
                storageClass: standard-csi
                pvcStorageRequest: 50G
Enabling ElasticSearch as a storage back end for events

To enable ElasticSearch as a storage back end for events, you must configure the ServiceTelemetry object.

Procedure
  • Configure the ServiceTelemetry object:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        events:
          elasticsearch:
            enabled: true
Configuring persistent storage for ElasticSearch

Use the additional parameters defined in backends.events.elasticsearch.storage.persistent to configure persistent storage options for ElasticSearch, such as storage class and volume size.

Use storageClass to define the back end storage class. If you do not set this parameter, the Service Telemetry Operator uses the default storage class for the OpenShift cluster.

Use the pvcStorageRequest parameter to define the minimum required volume size to satisfy the storage request. If volumes are statically defined, it is possible that a volume size larger than requested is used. By default, Service Telemetry Operator requests a volume size of 20Gi (20 Gibibytes).

Procedure
  • List the available storage classes:

    $ oc get storageclasses
    NAME                 PROVISIONER                RECLAIMPOLICY   VOLUMEBINDINGMODE      ALLOWVOLUMEEXPANSION   AGE
    csi-manila-ceph      manila.csi.openstack.org   Delete          Immediate              false                  20h
    standard (default)   kubernetes.io/cinder       Delete          WaitForFirstConsumer   true                   20h
    standard-csi         cinder.csi.openstack.org   Delete          WaitForFirstConsumer   true                   20h
  • Configure the ServiceTelemetry object:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        events:
          elasticsearch:
            enabled: true
            version: 7.16.1
            storage:
              strategy: persistent
              persistent:
                storageClass: standard-csi
                pvcStorageRequest: 50G
The clouds parameter

Use the clouds parameter to define which Smart Gateway objects deploy, thereby providing the interface for multiple monitored cloud environments to connect to an instance of STF. If a supporting back end is available, then metrics and events Smart Gateways for the default cloud configuration are created. By default, the Service Telemetry Operator creates Smart Gateways for cloud1.

You can create a list of cloud objects to control which Smart Gateways are created for the defined clouds. Each cloud consists of data types and collectors. Data types are metrics or events. Each data type consists of a list of collectors, the message bus subscription address, and a parameter to enable debugging. Available collectors for metrics are collectd, ceilometer, and sensubility. Available collectors for events are collectd and ceilometer. Ensure that the subscription address for each of these collectors is unique for every cloud, data type, and collector combination.

The default cloud1 configuration is represented by the following ServiceTelemetry object, which provides subscriptions and data storage of metrics and events for collectd, Ceilometer, and Sensubility data collectors for a particular cloud instance:

apiVersion: infra.watch/v1beta1
kind: ServiceTelemetry
metadata:
  name: stf-default
  namespace: service-telemetry
spec:
  clouds:
    - name: cloud1
      metrics:
        collectors:
          - collectorType: collectd
            subscriptionAddress: collectd/telemetry
          - collectorType: ceilometer
            subscriptionAddress: anycast/ceilometer/metering.sample
          - collectorType: sensubility
            subscriptionAddress: sensubility/telemetry
            debugEnabled: false
      events:
        collectors:
          - collectorType: collectd
            subscriptionAddress: collectd/notify
          - collectorType: ceilometer
            subscriptionAddress: anycast/ceilometer/event.sample

Each item of the clouds parameter represents a cloud instance. A cloud instance consists of three top-level parameters: name, metrics, and events. The metrics and events parameters represent the corresponding back end for storage of that data type. The collectors parameter specifies a list of objects made up of two required parameters, collectorType and subscriptionAddress, and these represent an instance of the Smart Gateway. The collectorType parameter specifies data collected by either collectd, Ceilometer, or Sensubility. The subscriptionAddress parameter provides the Apache Qpid Dispatch Router address to which a Smart Gateway subscribes.

You can use the optional Boolean parameter debugEnabled within the collectors parameter to enable additional console debugging in the running Smart Gateway pod.

Additional resources
The alerting parameter

Use the alerting parameter to control creation of an Alertmanager instance and the configuration of the storage back end. By default, alerting is enabled. For more information, see Alerts in Service Telemetry Framework.

The graphing parameter

Use the graphing parameter to control the creation of a Grafana instance. By default, graphing is disabled. For more information, see Dashboards in Service Telemetry Framework.

The highAvailability parameter

Use the highAvailability parameter to control the instantiation of multiple copies of STF components to reduce recovery time of components that fail or are rescheduled. By default, highAvailability is disabled. For more information, see High availability.

The transports parameter

Use the transports parameter to control the enablement of the message bus for a STF deployment. The only transport currently supported is Apache Qpid Dispatch Router. By default, the qdr transport is enabled.

Accessing user interfaces for STF components

In OpenShift, applications are exposed to the external network through a route. For more information about routes, see Configuring ingress cluster traffic.

In Service Telemetry Framework (STF), HTTPS routes are exposed for each service that has a web-based interface. These routes are protected by OpenShift RBAC and any user that has a ClusterRoleBinding that enables them to view OpenShift Namespaces can log in. For more information about RBAC, see Using RBAC to define and apply permissions.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. List the available web UI routes in the service-telemetry project:

    $ oc get routes | grep web
    default-alertmanager-proxy   default-alertmanager-proxy-service-telemetry.apps.infra.watch          default-alertmanager-proxy   web     reencrypt/Redirect   None
    default-prometheus-proxy     default-prometheus-proxy-service-telemetry.apps.infra.watch            default-prometheus-proxy     web     reencrypt/Redirect   None
  4. In a web browser, navigate to https://<route_address>; to access the web interface for the corresponding service.

Configuring an alternate observability strategy

To configure STF to skip the deployment of storage, visualization, and alerting backends, add observabilityStrategy: none to the ServiceTelemetry spec. In this mode, only Apache Qpid Dispatch Router routers and metrics Smart Gateways are deployed, and you must configure an external Prometheus-compatible system to collect metrics from the STF Smart Gateways.

Currently, only metrics are supported when you set observabilityStrategy to none. Events Smart Gateways are not deployed.
Procedure
  1. Create a ServiceTelemetry object with the property observabilityStrategy: none in the spec parameter. The manifest shows results in a default deployment of STF that is suitable for receiving telemetry from a single cloud with all metrics collector types.

    $ oc apply -f - <<EOF
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      observabilityStrategy: none
    EOF
  2. To verify that all workloads are operating correctly, view the pods and the status of each pod:

    $ oc get pods
    NAME                                                      READY   STATUS    RESTARTS   AGE
    default-cloud1-ceil-meter-smartgateway-59c845d65b-gzhcs   3/3     Running   0          132m
    default-cloud1-coll-meter-smartgateway-75bbd948b9-d5phm   3/3     Running   0          132m
    default-cloud1-sens-meter-smartgateway-7fdbb57b6d-dh2g9   3/3     Running   0          132m
    default-interconnect-668d5bbcd6-57b2l                     1/1     Running   0          132m
    interconnect-operator-b8f5bb647-tlp5t                     1/1     Running   0          47h
    service-telemetry-operator-566b9dd695-wkvjq               1/1     Running   0          156m
    smart-gateway-operator-58d77dcf7-6xsq7                    1/1     Running   0          47h
Additional resources

For more information about configuring additional clouds or to change the set of supported collectors, see Deploying Smart Gateways

Removing Service Telemetry Framework from the OpenShift environment

Remove Service Telemetry Framework (STF) from an OpenShift environment if you no longer require the STF functionality.

Deleting the namespace

To remove the operational resources for STF from OpenShift, delete the namespace.

Procedure
  1. Run the oc delete command:

    $ oc delete project service-telemetry
  2. Verify that the resources have been deleted from the namespace:

    $ oc get all
    No resources found.

Removing the CatalogSource

If you do not expect to install Service Telemetry Framework (STF) again, delete the CatalogSource. When you remove the CatalogSource, PackageManifests related to STF are automatically removed from the Operator Lifecycle Manager catalog.

Procedure
  1. Delete the CatalogSource:

    $ oc delete --namespace=openshift-marketplace catalogsource infrawatch-operators
    catalogsource.operators.coreos.com "infrawatch-operators" deleted
  2. Verify that the STF PackageManifests are removed from the platform. If successful, the following command returns no result:

    $ oc get packagemanifests | grep InfraWatch
  3. If you enabled the OperatorHub.io Community Catalog Source during the installation process and you no longer need this catalog source, delete it:

    $ oc delete --namespace=openshift-marketplace catalogsource operatorhubio-operators
    catalogsource.operators.coreos.com "operatorhubio-operators" deleted
Additional resources

For more information about the OperatorHub.io Community Catalog Source, see Deploying Service Telemetry Framework to the OpenShift environment.

Configuring OpenStack for Service Telemetry Framework

To collect metrics, events, or both, and to send them to the Service Telemetry Framework (STF) storage domain, you must configure the OpenStack (OSP) overcloud to enable data collection and transport.

STF can support both single and multiple clouds. The default configuration in OSP and STF set up for a single cloud installation.

Deploying OpenStack overcloud for Service Telemetry Framework

As part of the OpenStack (OSP) overcloud deployment, you must configure the data collectors and the data transport to Service Telemetry Framework (STF).

Additional resources

Retrieving the Apache Qpid Dispatch Router route address

When you configure the OpenStack (OSP) overcloud for Service Telemetry Framework (STF), you must provide the Apache Qpid Dispatch Router route address in the STF connection file.

Procedure
  1. Log in to your OpenShift environment.

  2. In the service-telemetry project, retrieve the Apache Qpid Dispatch Router route address:

    $ oc get routes -ogo-template='{{ range .items }}{{printf "%s\n" .spec.host }}{{ end }}' | grep "\-5671"
    default-interconnect-5671-service-telemetry.apps.infra.watch

Creating the base configuration for STF

To configure the base parameters to provide a compatible data collection and transport for Service Telemetry Framework (STF), you must create a file that defines the default data collection values.

Procedure
  1. Log in to the OpenStack (OSP) undercloud as the stack user.

  2. Create a configuration file called enable-stf.yaml in the /home/stack directory.

    Setting EventPipelinePublishers and PipelinePublishers to empty lists results in no event or metric data passing to OSP telemetry components, such as Gnocchi or Panko. If you need to send data to additional pipelines, the Ceilometer polling interval of 30 seconds, as specified in ExtraConfig, might overwhelm the OSP telemetry components, and you must increase the interval to a larger value, such as 300. Increasing the value to a longer polling interval results in less telemetry resolution in STF.

    To enable collection of telemetry with STF and Gnocchi, see Sending metrics to Gnocchi and Service Telemetry Framework

enable-stf.yaml
parameter_defaults:
    # only send to STF, not other publishers
    EventPipelinePublishers: []
    PipelinePublishers: []

    # manage the polling and pipeline configuration files for Ceilometer agents
    ManagePolling: true
    ManagePipeline: true

    # enable Ceilometer metrics and events
    CeilometerQdrPublishMetrics: true
    CeilometerQdrPublishEvents: true

    # enable collection of API status
    CollectdEnableSensubility: true
    CollectdSensubilityTransport: amqp1

    # enable collection of containerized service metrics
    CollectdEnableLibpodstats: true

    # set collectd overrides for higher telemetry resolution and extra plugins
    # to load
    CollectdConnectionType: amqp1
    CollectdAmqpInterval: 5
    CollectdDefaultPollingInterval: 5
    CollectdExtraPlugins:
    - vmem

    # set standard prefixes for where metrics and events are published to QDR
    MetricsQdrAddresses:
    - prefix: 'collectd'
      distribution: multicast
    - prefix: 'anycast/ceilometer'
      distribution: multicast

    ExtraConfig:
        ceilometer::agent::polling::polling_interval: 30
        ceilometer::agent::polling::polling_meters:
        - cpu
        - disk.*
        - ip.*
        - image.*
        - memory
        - memory.*
        - network.*
        - perf.*
        - port
        - port.*
        - switch
        - switch.*
        - storage.*
        - volume.*

        # to avoid filling the memory buffers if disconnected from the message bus
        # note: this may need an adjustment if there are many metrics to be sent.
        collectd::plugin::amqp1::send_queue_limit: 5000

        # receive extra information about virtual memory
        collectd::plugin::vmem::verbose: true

        # provide name and uuid in addition to hostname for better correlation
        # to ceilometer data
        collectd::plugin::virt::hostname_format: "name uuid hostname"

        # provide the human-friendly name of the virtual instance
        collectd::plugin::virt::plugin_instance_format: metadata

        # set memcached collectd plugin to report its metrics by hostname
        # rather than host IP, ensuring metrics in the dashboard remain uniform
        collectd::plugin::memcached::instances:
          local:
            host: "%{hiera('fqdn_canonical')}"
            port: 11211

Configuring the STF connection for the overcloud

To configure the Service Telemetry Framework (STF) connection, you must create a file that contains the connection configuration of the Apache Qpid Dispatch Router for the overcloud to the STF deployment. Enable the collection of events and storage of the events in STF and deploy the overcloud. The default configuration is for a single cloud instance with the default message bus topics. For configuration of multiple cloud deployments, see Configuring multiple clouds.

Prerequisites
Procedure
  1. Log in to the OSP undercloud as the stack user.

  2. Create a configuration file called stf-connectors.yaml in the /home/stack directory.

  3. In the stf-connectors.yaml file, configure the MetricsQdrConnectors address to connect the Apache Qpid Dispatch Router on the overcloud to the STF deployment. You configure the topic addresses for Sensubility, Ceilometer, and collectd in this file to match the defaults in STF. For more information about customizing topics and cloud configuration, see Configuring multiple clouds.

    stf-connectors.yaml
    resource_registry:
      OS::TripleO::Services::Collectd: /usr/share/openstack-tripleo-heat-templates/deployment/metrics/collectd-container-puppet.yaml    (1)
    
    parameter_defaults:
        MetricsQdrConnectors:
            - host: stf-default-interconnect-5671-service-telemetry.apps.infra.watch   (2)
              port: 443
              role: edge
              verifyHostname: false
              sslProfile: sslProfile
    
        MetricsQdrSSLProfiles:
            - name: sslProfile
    
        CeilometerQdrEventsConfig:
            driver: amqp
            topic: cloud1-event   (3)
    
        CeilometerQdrMetricsConfig:
            driver: amqp
            topic: cloud1-metering   (4)
    
        CollectdAmqpInstances:
            cloud1-notify:        (5)
                notify: true
                format: JSON
                presettle: false
            cloud1-telemetry:     (6)
                format: JSON
                presettle: false
    
        CollectdSensubilityResultsChannel: sensubility/cloud1-telemetry (7)
    1 Directly load the collectd service because you are not including the collectd-write-qdr.yaml environment file for multiple cloud deployments.
    2 Replace the host parameter with the value of HOST/PORT that you retrieved in Retrieving the Apache Qpid Dispatch Router route address.
    3 Define the topic for Ceilometer events. The format of this value is anycast/ceilometer/cloud1-event.sample.
    4 Define the topic for Ceilometer metrics. The format of this value is`anycast/ceilometer/cloud1-metering.sample`.
    5 Define the topic for collectd events. The format of this value is collectd/cloud1-notify.
    6 Define the topic for collectd metrics. The format of this value is collectd/cloud1-telemetry.
    7 Define the topic for collectd-sensubility events. The value is the exact string sensubility/cloud1-telemetry.

Deploying the overcloud

Deploy or update the overcloud with the required environment files so that data is collected and transmitted to Service Telemetry Framework (STF).

Procedure
  1. Log in to the OpenStack (OSP) undercloud as the stack user.

  2. Source the authentication file:

    [stack@undercloud-0 ~]$ source stackrc
    
    (undercloud) [stack@undercloud-0 ~]$
  3. Add the following files to your OSP TripleO deployment to configure data collection and Apache Qpid Dispatch Router:

    • The ceilometer-write-qdr.yaml file to ensure that Ceilometer telemetry and events are sent to STF

    • The qdr-edge-only.yaml file to ensure that the message bus is enabled and connected to STF message bus routers

    • The enable-stf.yaml environment file to ensure defaults are configured correctly

    • The stf-connectors.yaml environment file to define the connection to STF

  4. Deploy the OSP overcloud:

    (undercloud) [stack@undercloud-0 ~]$ openstack overcloud deploy <other_arguments>
    --templates /usr/share/openstack-tripleo-heat-templates \
      --environment-file <...other_environment_files...> \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/ceilometer-write-qdr.yaml \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/qdr-edge-only.yaml \
      --environment-file /home/stack/enable-stf.yaml \
      --environment-file /home/stack/stf-connectors.yaml

Validating client-side installation

To validate data collection from the Service Telemetry Framework (STF) storage domain, query the data sources for delivered data. To validate individual nodes in the OpenStack (OSP) deployment, use SSH to connect to the console.

Some telemetry data is available only when OSP has active workloads.
Procedure
  1. Log in to an overcloud node, for example, controller-0.

  2. Ensure that the metrics_qdr container is running on the node:

    $ sudo podman container inspect --format '{{.State.Status}}' metrics_qdr
    
    running
  3. Return the internal network address on which Apache Qpid Dispatch Router is running, for example, 172.17.1.44 listening on port 5666:

    $ sudo podman exec -it metrics_qdr cat /etc/qpid-dispatch/qdrouterd.conf
    
    listener {
        host: 172.17.1.44
        port: 5666
        authenticatePeer: no
        saslMechanisms: ANONYMOUS
    }
  4. Return a list of connections to the local Apache Qpid Dispatch Router:

    $ sudo podman exec -it metrics_qdr qdstat --bus=172.17.1.44:5666 --connections
    
    Connections
      id   host                                                                  container                                                                                                  role    dir  security                            authentication  tenant
      ============================================================================================================================================================================================================================================================================================
      1    default-interconnect-5671-service-telemetry.apps.infra.watch:443      default-interconnect-7458fd4d69-bgzfb                                                                      edge    out  TLSv1.2(DHE-RSA-AES256-GCM-SHA384)  anonymous-user
      12   172.17.1.44:60290                                                     openstack.org/om/container/controller-0/ceilometer-agent-notification/25/5c02cee550f143ec9ea030db5cccba14  normal  in   no-security                         no-auth
      16   172.17.1.44:36408                                                     metrics                                                                                                    normal  in   no-security                         anonymous-user
      899  172.17.1.44:39500                                                     10a2e99d-1b8a-4329-b48c-4335e5f75c84                                                                       normal  in   no-security                         no-auth

    There are four connections:

    • Outbound connection to STF

    • Inbound connection from ceilometer

    • Inbound connection from collectd

    • Inbound connection from our qdstat client

      The outbound STF connection is provided to the MetricsQdrConnectors host parameter and is the route for the STF storage domain. The other hosts are internal network addresses of the client connections to this Apache Qpid Dispatch Router.

  5. To ensure that messages are delivered, list the links, and view the _edge address in the deliv column for delivery of messages:

    $ sudo podman exec -it metrics_qdr qdstat --bus=172.17.1.44:5666 --links
    Router Links
      type      dir  conn id  id    peer  class   addr                  phs  cap  pri  undel  unsett  deliv    presett  psdrop  acc  rej  rel     mod  delay  rate
      ===========================================================================================================================================================
      endpoint  out  1        5           local   _edge                      250  0    0      0       2979926  0        0       0    0    2979926 0    0      0
      endpoint  in   1        6                                              250  0    0      0       0        0        0       0    0    0       0    0      0
      endpoint  in   1        7                                              250  0    0      0       0        0        0       0    0    0       0    0      0
      endpoint  out  1        8                                              250  0    0      0       0        0        0       0    0    0       0    0      0
      endpoint  in   1        9                                              250  0    0      0       0        0        0       0    0    0       0    0      0
      endpoint  out  1        10                                             250  0    0      0       911      911      0       0    0    0       0    911    0
      endpoint  in   1        11                                             250  0    0      0       0        911      0       0    0    0       0    0      0
      endpoint  out  12       32          local   temp.lSY6Mcicol4J2Kp       250  0    0      0       0        0        0       0    0    0       0    0      0
      endpoint  in   16       41                                             250  0    0      0       2979924  0        0       0    0    2979924 0    0      0
      endpoint  in   912      1834        mobile  $management           0    250  0    0      0       1        0        0       1    0    0       0    0      0
      endpoint  out  912      1835        local   temp.9Ok2resI9tmt+CT       250  0    0      0       0        0        0       0    0    0       0    0      0
  6. To list the addresses from OSP nodes to STF, connect to OpenShift to retrieve the Apache Qpid Dispatch Router pod name and list the connections. List the available Apache Qpid Dispatch Router pods:

    $ oc get pods -l application=default-interconnect
    
    NAME                                    READY   STATUS    RESTARTS   AGE
    default-interconnect-7458fd4d69-bgzfb   1/1     Running   0          6d21h
  7. Connect to the pod and list the known connections. In this example, there are three edge connections from the OSP nodes with connection id 22, 23, and 24:

    $ oc exec -it default-interconnect-7458fd4d69-bgzfb -- qdstat --connections
    
    2020-04-21 18:25:47.243852 UTC
    default-interconnect-7458fd4d69-bgzfb
    
    Connections
      id  host               container                                                      role    dir  security                                authentication  tenant  last dlv      uptime
      ===============================================================================================================================================================================================
      5   10.129.0.110:48498  bridge-3f5                                                    edge    in   no-security                             anonymous-user          000:00:00:02  000:17:36:29
      6   10.129.0.111:43254  rcv[default-cloud1-ceil-meter-smartgateway-58f885c76d-xmxwn]  edge    in   no-security                             anonymous-user          000:00:00:02  000:17:36:20
      7   10.130.0.109:50518  rcv[default-cloud1-coll-event-smartgateway-58fbbd4485-rl9bd]  normal  in   no-security                             anonymous-user          -             000:17:36:11
      8   10.130.0.110:33802  rcv[default-cloud1-ceil-event-smartgateway-6cfb65478c-g5q82]  normal  in   no-security                             anonymous-user          000:01:26:18  000:17:36:05
      22  10.128.0.1:51948   Router.ceph-0.redhat.local                                     edge    in   TLSv1/SSLv3(DHE-RSA-AES256-GCM-SHA384)  anonymous-user          000:00:00:03  000:22:08:43
      23  10.128.0.1:51950   Router.compute-0.redhat.local                                  edge    in   TLSv1/SSLv3(DHE-RSA-AES256-GCM-SHA384)  anonymous-user          000:00:00:03  000:22:08:43
      24  10.128.0.1:52082   Router.controller-0.redhat.local                               edge    in   TLSv1/SSLv3(DHE-RSA-AES256-GCM-SHA384)  anonymous-user          000:00:00:00  000:22:08:34
      27  127.0.0.1:42202    c2f541c1-4c97-4b37-a189-a396c08fb079                           normal  in   no-security                             no-auth                 000:00:00:00  000:00:00:00
  8. To view the number of messages delivered by the network, use each address with the oc exec command:

    $ oc exec -it default-interconnect-7458fd4d69-bgzfb -- qdstat --address
    
    2020-04-21 18:20:10.293258 UTC
    default-interconnect-7458fd4d69-bgzfb
    
    Router Addresses
      class   addr                                phs  distrib    pri  local  remote  in           out          thru  fallback
      ==========================================================================================================================
      mobile  anycast/ceilometer/event.sample     0    balanced   -    1      0       970          970          0     0
      mobile  anycast/ceilometer/metering.sample  0    balanced   -    1      0       2,344,833    2,344,833    0     0
      mobile  collectd/notify                     0    multicast  -    1      0       70           70           0     0
      mobile  collectd/telemetry                  0    multicast  -    1      0       216,128,890  216,128,890  0     0

Sending metrics to Gnocchi and Service Telemetry Framework

To send metrics to Service Telemetry Framework (STF) and Gnocchi simultaneously, you must include an environment file in your deployment to enable an additional publisher.

If you need to send data to additional pipelines, the Ceilometer polling interval of 30 seconds, as specified in ExtraConfig, might overwhelm the OSP telemetry components, and you must increase the interval to a larger value, such as 300. Increasing the value to a longer polling interval results in less telemetry resolution in STF.
Prerequisites
Procedure
  1. Create an environment file named gnocchi-connectors.yaml in the /home/stack directory.

    resource_registry:
        OS::TripleO::Services::GnocchiApi: /usr/share/openstack-tripleo-heat-templates/deployment/gnocchi/gnocchi-api-container-puppet.yaml
        OS::TripleO::Services::GnocchiMetricd: /usr/share/openstack-tripleo-heat-templates/deployment/gnocchi/gnocchi-metricd-container-puppet.yaml
        OS::TripleO::Services::GnocchiStatsd: /usr/share/openstack-tripleo-heat-templates/deployment/gnocchi/gnocchi-statsd-container-puppet.yaml
        OS::TripleO::Services::AodhApi: /usr/share/openstack-tripleo-heat-templates/deployment/aodh/aodh-api-container-puppet.yaml
        OS::TripleO::Services::AodhEvaluator: /usr/share/openstack-tripleo-heat-templates/deployment/aodh/aodh-evaluator-container-puppet.yaml
        OS::TripleO::Services::AodhNotifier: /usr/share/openstack-tripleo-heat-templates/deployment/aodh/aodh-notifier-container-puppet.yaml
        OS::TripleO::Services::AodhListener: /usr/share/openstack-tripleo-heat-templates/deployment/aodh/aodh-listener-container-puppet.yaml
    
    parameter_defaults:
        CeilometerEnableGnocchi: true
        CeilometerEnablePanko: false
        GnocchiArchivePolicy: 'high'
        GnocchiBackend: 'rbd'
        GnocchiRbdPoolName: 'metrics'
    
        EventPipelinePublishers: ['gnocchi://?filter_project=service']
        PipelinePublishers: ['gnocchi://?filter_project=service']
  2. Add the environment file gnocchi-connectors.yaml to the deployment command. Replace <other_arguments> with files that are applicable to your environment.

    $ openstack overcloud deploy <other_arguments>
    --templates /usr/share/openstack-tripleo-heat-templates \
      --environment-file <...other_environment_files...> \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/ceilometer-write-qdr.yaml \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/collectd-write-qdr.yaml \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/qdr-edge-only.yaml \
      --environment-file /home/stack/enable-stf.yaml \
      --environment-file /home/stack/stf-connectors.yaml \
      --environment-file /home/stack/gnocchi-connectors.yaml
  3. To ensure that the configuration was successful, verify the content of the file /var/lib/config-data/puppet-generated/ceilometer/etc/ceilometer/pipeline.yaml on a Controller node. Ensure that the publishers section of the file contains information for both notifier and Gnocchi.

    sources:
        - name: meter_source
          meters:
              - "*"
          sinks:
              - meter_sink
    sinks:
        - name: meter_sink
          publishers:
              - gnocchi://?filter_project=service
              - notifier://172.17.1.35:5666/?driver=amqp&topic=metering

Deploying to non-standard network topologies

If your nodes are on a separate network from the default InternalApi network, you must make configuration adjustments so that Apache Qpid Dispatch Router can transport data to the Service Telemetry Framework (STF) server instance. This scenario is typical in a spine-leaf or a DCN topology. For more information about DCN configuration, see the Spine Leaf Networking guide.

If you use STF with OpenStack (OSP) Train and plan to monitor your Ceph, Block, or Object Storage nodes, you must make configuration changes that are similar to the configuration changes that you make to the spine-leaf and DCN network configuration. To monitor Ceph nodes, use the CephStorageExtraConfig parameter to define which network interface to load into the Apache Qpid Dispatch Router and collectd configuration files.

CephStorageExtraConfig:
  tripleo::profile::base::metrics::collectd::amqp_host: "%{hiera('storage')}"
  tripleo::profile::base::metrics::qdr::listener_addr: "%{hiera('storage')}"
  tripleo::profile::base::ceilometer::agent::notification::notifier_host_addr: "%{hiera('storage')}"

Similarly, you must specify BlockStorageExtraConfig and ObjectStorageExtraConfig parameters if your environment uses Block and Object Storage roles.

To deploy a spine-leaf topology, you must create roles and networks, then assign those networks to the available roles. When you configure data collection and transport for STF for an OSP deployment, the default network for roles is InternalApi. For Ceph, Block and Object storage roles, the default network is Storage. Because a spine-leaf configuration can result in different networks being assigned to different Leaf groupings and those names are typically unique, additional configuration is required in the parameter_defaults section of the OSP environment files.

Procedure
  1. Document which networks are available for each of the Leaf roles. For examples of network name definitions, see Creating a network data file in the Spine Leaf Networking guide. For more information about the creation of the Leaf groupings (roles) and assignment of the networks to those groupings, see Creating a roles data file in the Spine Leaf Networking guide.

  2. Add the following configuration example to the ExtraConfig section for each of the leaf roles. In this example, internal_api_subnet is the value defined in the name_lower parameter of your network definition (with _subnet appended to the name for Leaf 0) , and is the network to which the ComputeLeaf0 leaf role is connected. In this case, the network identification of 0 corresponds to the Compute role for leaf 0, and represents a value that is different from the default internal API network name.

    For the ComputeLeaf0 leaf role, specify extra configuration to perform a hiera lookup to determine which network interface for a particular network to assign to the collectd AMQP host parameter. Perform the same configuration for the Apache Qpid Dispatch Router listener address parameter.

    ComputeLeaf0ExtraConfig:
      tripleo::profile::base::metrics::collectd::amqp_host: "%{hiera('internal_api_subnet')}"
      tripleo::profile::base::metrics::qdr::listener_addr: "%{hiera('internal_api_subnet')}"

    Additional leaf roles typically replace _subnet with _leafN. N represents a unique identifier for the leaf.

    ComputeLeaf1ExtraConfig:
      tripleo::profile::base::metrics::collectd::amqp_host: "%{hiera('internal_api_leaf1')}"
      tripleo::profile::base::metrics::qdr::listener_addr: "%{hiera('internal_api_leaf1')}"

    This example configuration is on a CephStorage leaf role:

    CephStorageLeaf0ExtraConfig:
      tripleo::profile::base::metrics::collectd::amqp_host: "%{hiera('storage_subnet')}"
      tripleo::profile::base::metrics::qdr::listener_addr: "%{hiera('storage_subnet')}"

Configuring multiple clouds

You can configure multiple OpenStack (OSP) clouds to target a single instance of Service Telemetry Framework (STF). When you configure multiple clouds, every cloud must send metrics and events on their own unique message bus topic. In the STF deployment, Smart Gateway instances listen on these topics to save information to the common data store. Data that is stored by the Smart Gateway in the data storage domain is filtered by using the metadata that each of Smart Gateways creates.

An example of two OSP clouds connecting to STF
Figure 3. Two OSP clouds connect to STF

To configure the OSP overcloud for a multiple cloud scenario, complete the following tasks:

  1. Plan the AMQP address prefixes that you want to use for each cloud. For more information, see Planning AMQP address prefixes.

  2. Deploy metrics and events consumer Smart Gateways for each cloud to listen on the corresponding address prefixes. For more information, see Deploying Smart Gateways.

  3. Configure each cloud with a unique domain name. For more information, see Setting a unique cloud domain.

  4. Create the base configuration for STF. For more information, see Creating the base configuration for STF.

  5. Configure each cloud to send its metrics and events to STF on the correct address. For more information, see Creating the OpenStack environment file for multiple clouds.

Planning AMQP address prefixes

By default, OpenStack (OSP) nodes receive data through two data collectors; collectd and Ceilometer. The collectd-sensubility plugin requires a unique address. These components send telemetry data or notifications to the respective AMQP addresses, for example, collectd/telemetry. STF Smart Gateways listen on those AMQP addresses for data. To support multiple clouds and to identify which cloud generated the monitoring data, configure each cloud to send data to a unique address. Add a cloud identifier prefix to the second part of the address. The following list shows some example addresses and identifiers:

  • collectd/cloud1-telemetry

  • collectd/cloud1-notify

  • sensubility/cloud1-telemetry

  • anycast/ceilometer/cloud1-metering.sample

  • anycast/ceilometer/cloud1-event.sample

  • collectd/cloud2-telemetry

  • collectd/cloud2-notify

  • sensubility/cloud2-telemetry

  • anycast/ceilometer/cloud2-metering.sample

  • anycast/ceilometer/cloud2-event.sample

  • collectd/us-east-1-telemetry

  • collectd/us-west-3-telemetry

Deploying Smart Gateways

You must deploy a Smart Gateway for each of the data collection types for each cloud; one for collectd metrics, one for collectd events, one for Ceilometer metrics, one for Ceilometer events, and one for collectd-sensubility metrics. Configure each of the Smart Gateways to listen on the AMQP address that you define for the corresponding cloud. To define Smart Gateways, configure the clouds parameter in the ServiceTelemetry manifest.

When you deploy STF for the first time, Smart Gateway manifests are created that define the initial Smart Gateways for a single cloud. When you deploy Smart Gateways for multiple cloud support, you deploy multiple Smart Gateways for each of the data collection types that handle the metrics and the events data for each cloud. The initial Smart Gateways are defined in cloud1 with the following subscription addresses:

collector

type

default subscription address

collectd

metrics

collectd/telemetry

collectd

events

collectd/notify

collectd-sensubility

metrics

sensubility/telemetry

Ceilometer

metrics

anycast/ceilometer/metering.sample

Ceilometer

events

anycast/ceilometer/event.sample

Prerequisites
  • You have determined your cloud naming scheme. For more information about determining your naming scheme, see Planning AMQP address prefixes.

  • You have created your list of clouds objects. For more information about creating the content for the clouds parameter, see The clouds parameter.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Edit the default ServiceTelemetry object and add a clouds parameter with your configuration:

    Long cloud names might exceed the maximum pod name of 63 characters. Ensure that the combination of the ServiceTelemetry name default and the clouds.name does not exceed 19 characters. Cloud names cannot contain any special characters, such as -. Limit cloud names to alphanumeric (a-z, 0-9).

    Topic addresses have no character limitation and can be different from the clouds.name value.

    $ oc edit stf default
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      ...
    spec:
      ...
      clouds:
      - name: cloud1
        events:
          collectors:
          - collectorType: collectd
            subscriptionAddress: collectd/cloud1-notify
          - collectorType: ceilometer
            subscriptionAddress: anycast/ceilometer/cloud1-event.sample
        metrics:
          collectors:
          - collectorType: collectd
            subscriptionAddress: collectd/cloud1-telemetry
          - collectorType: sensubility
            subscriptionAddress: sensubility/cloud1-telemetry
          - collectorType: ceilometer
            subscriptionAddress: anycast/ceilometer/cloud1-metering.sample
      - name: cloud2
        events:
          ...
  4. Save the ServiceTelemetry object.

  5. Verify that each Smart Gateway is running. This can take several minutes depending on the number of Smart Gateways:

    $ oc get po -l app=smart-gateway
    NAME                                                      READY   STATUS    RESTARTS   AGE
    default-cloud1-ceil-event-smartgateway-6cfb65478c-g5q82   2/2     Running   0          13h
    default-cloud1-ceil-meter-smartgateway-58f885c76d-xmxwn   2/2     Running   0          13h
    default-cloud1-coll-event-smartgateway-58fbbd4485-rl9bd   2/2     Running   0          13h
    default-cloud1-coll-meter-smartgateway-7c6fc495c4-jn728   2/2     Running   0          13h
    default-cloud1-sens-meter-smartgateway-8h4tc445a2-mm683   2/2     Running   0          13h

Deleting the default Smart Gateways

After you configure Service Telemetry Framework (STF) for multiple clouds, you can delete the default Smart Gateways if they are no longer in use. The Service Telemetry Operator can remove SmartGateway objects that were created but are no longer listed in the ServiceTelemetry clouds list of objects. To enable the removal of SmartGateway objects that are not defined by the clouds parameter, you must set the cloudsRemoveOnMissing parameter to true in the ServiceTelemetry manifest.

If you do not want to deploy any Smart Gateways, define an empty clouds list by using the clouds: [] parameter.
The cloudsRemoveOnMissing parameter is disabled by default. If you enable the cloudsRemoveOnMissing parameter, you remove any manually created SmartGateway objects in the current namespace without any possibility to restore.
Procedure
  1. Define your clouds parameter with the list of cloud objects that you want the Service Telemetry Operator to manage. For more information, see The clouds parameter.

  2. Edit the ServiceTelemetry object and add the cloudsRemoveOnMissing parameter:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      ...
    spec:
      ...
      cloudsRemoveOnMissing: true
      clouds:
        ...
  3. Save the modifications.

  4. Verify that the Operator deleted the Smart Gateways. This can take several minutes while the Operators reconcile the changes:

    $ oc get smartgateways

Setting a unique cloud domain

To ensure that Apache Qpid Dispatch Router router connections from OpenStack (OSP) to Service Telemetry Framework (STF) are unique and do not conflict, configure the CloudDomain parameter.

Procedure
  1. Create a new environment file, for example, hostnames.yaml.

  2. Set the CloudDomain parameter in the environment file, as shown in the following example:

    hostnames.yaml
    parameter_defaults:
        CloudDomain: newyork-west-04
        CephStorageHostnameFormat: 'ceph-%index%'
        ObjectStorageHostnameFormat: 'swift-%index%'
        ComputeHostnameFormat: 'compute-%index%'
  3. Add the new environment file to your deployment. For more information, see Creating the OpenStack environment file for multiple clouds and Core overcloud parameters in the Overcloud Parameters guide.

Creating the OpenStack environment file for multiple clouds

To label traffic according to the cloud of origin, you must create a configuration with cloud-specific instance names. Create an stf-connectors.yaml file and adjust the values of CeilometerQdrEventsConfig, CeilometerQdrMetricsConfig and CollectdAmqpInstances to match the AMQP address prefix scheme.

If you enabled container health and API status monitoring, you must also modify the CollectdSensubilityResultsChannel parameter. For more information, see OpenStack API status and containerized services health.
Prerequisites
Procedure
  1. Log in to the OpenStack undercloud as the stack user.

  2. Create a configuration file called stf-connectors.yaml in the /home/stack directory.

  3. In the stf-connectors.yaml file, configure the MetricsQdrConnectors address to connect to the Apache Qpid Dispatch Router on the overcloud deployment. Configure the CeilometerQdrEventsConfig, CeilometerQdrMetricsConfig, CollectdAmqpInstances, and CollectdSensubilityResultsChannel topic values to match the AMQP address that you want for this cloud deployment.

    stf-connectors.yaml
    resource_registry:
      OS::TripleO::Services::Collectd: /usr/share/openstack-tripleo-heat-templates/deployment/metrics/collectd-container-puppet.yaml    (1)
    
    parameter_defaults:
        MetricsQdrConnectors:
            - host: stf-default-interconnect-5671-service-telemetry.apps.infra.watch   (2)
              port: 443
              role: edge
              verifyHostname: false
              sslProfile: sslProfile
    
        MetricsQdrSSLProfiles:
            - name: sslProfile
    
        CeilometerQdrEventsConfig:
            driver: amqp
            topic: cloud1-event   (3)
    
        CeilometerQdrMetricsConfig:
            driver: amqp
            topic: cloud1-metering   (4)
    
        CollectdAmqpInstances:
            cloud1-notify:        (5)
                notify: true
                format: JSON
                presettle: false
            cloud1-telemetry:     (6)
                format: JSON
                presettle: false
    
        CollectdSensubilityResultsChannel: sensubility/cloud1-telemetry (7)
    1 Directly load the collectd service because you are not including the collectd-write-qdr.yaml environment file for multiple cloud deployments.
    2 Replace the host parameter with the value of HOST/PORT that you retrieved in Retrieving the Apache Qpid Dispatch Router route address.
    3 Define the topic for Ceilometer events. This value is the address format of anycast/ceilometer/cloud1-event.sample.
    4 Define the topic for Ceilometer metrics. This value is the address format of anycast/ceilometer/cloud1-metering.sample.
    5 Define the topic for collectd events. This value is the format of collectd/cloud1-notify.
    6 Define the topic for collectd metrics. This value is the format of collectd/cloud1-telemetry.
    7 Define the topic for collectd-sensubility events. Ensure that this value is the exact string format sensubility/cloud1-telemetry
  4. Ensure that the naming convention in the stf-connectors.yaml file aligns with the spec.bridge.amqpUrl field in the Smart Gateway configuration. For example, configure the CeilometerQdrEventsConfig.topic field to a value of cloud1-event.

  5. Source the authentication file:

    [stack@undercloud-0 ~]$ source stackrc
    
    (undercloud) [stack@undercloud-0 ~]$
  6. Include the stf-connectors.yaml file and unique domain name environment file hostnames.yaml in the openstack overcloud deployment command, with any other environment files relevant to your environment:

    If you use the collectd-write-qdr.yaml file with a custom CollectdAmqpInstances parameter, data publishes to the custom and default topics. In a multiple cloud environment, the configuration of the resource_registry parameter in the stf-connectors.yaml file loads the collectd service.
    (undercloud) [stack@undercloud-0 ~]$ openstack overcloud deploy <other_arguments>
    --templates /usr/share/openstack-tripleo-heat-templates \
      --environment-file <...other_environment_files...> \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/ceilometer-write-qdr.yaml \
      --environment-file /usr/share/openstack-tripleo-heat-templates/environments/metrics/qdr-edge-only.yaml \
      --environment-file /home/stack/hostnames.yaml \
      --environment-file /home/stack/enable-stf.yaml \
      --environment-file /home/stack/stf-connectors.yaml
  7. Deploy the OpenStack overcloud.

Additional resources

Querying metrics data from multiple clouds

Data stored in Prometheus has a service label according to the Smart Gateway it was scraped from. You can use this label to query data from a specific cloud.

To query data from a specific cloud, use a Prometheus promql query that matches the associated service label; for example: collectd_uptime{service="default-cloud1-coll-meter"}.

Using operational features of Service Telemetry Framework

You can use the following operational features to provide additional functionality to the Service Telemetry Framework (STF):

Dashboards in Service Telemetry Framework

Use the third-party application, Grafana, to visualize system-level metrics that collectd and Ceilometer gathers for each individual host node.

For more information about configuring collectd, see Deploying OpenStack overcloud for Service Telemetry Framework.

You can use two dashboards to monitor a cloud:

Infrastructure dashboard

Use the infrastructure dashboard to view metrics for a single node at a time. Select a node from the upper left corner of the dashboard.

Cloud view dashboard

Use the cloud view dashboard to view panels to monitor service resource usage, API stats, and cloud events. You must enable API health monitoring and service monitoring to provide the data for this dashboard. API health monitoring is enabled by default in the STF base configuration. For more information, see Creating the base configuration for STF.

Configuring Grafana to host the dashboard

Grafana is not included in the default Service Telemetry Framework (STF) deployment so you must deploy the Grafana Operator from OperatorHub.io. When you use the Service Telemetry Operator to deploy Grafana, it results in a Grafana instance and the configuration of the default data sources for the local STF deployment.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Deploy the Grafana operator:

    $ oc apply -f - <<EOF
    apiVersion: operators.coreos.com/v1alpha1
    kind: Subscription
    metadata:
      name: grafana-operator
      namespace: service-telemetry
    spec:
      channel: alpha
      installPlanApproval: Automatic
      name: grafana-operator
      source: operatorhubio-operators
      sourceNamespace: openshift-marketplace
    EOF
  4. Verify that the Operator launched successfully. In the command output, if the value of the PHASE column is Succeeded, the Operator launched successfully:

    $ oc get csv --selector operators.coreos.com/grafana-operator.service-telemetry
    
    NAME                       DISPLAY            VERSION   REPLACES                   PHASE
    grafana-operator.v3.10.3   Grafana Operator   3.10.3    grafana-operator.v3.10.2   Succeeded
  5. To launch a Grafana instance, create or modify the ServiceTelemetry object. Set graphing.enabled and graphing.grafana.ingressEnabled to true:

    $ oc edit stf default
    
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    ...
    spec:
      ...
      graphing:
        enabled: true
        grafana:
          ingressEnabled: true
  6. Verify that the Grafana instance deployed:

    $ oc get pod -l app=grafana
    
    NAME                                  READY   STATUS    RESTARTS   AGE
    grafana-deployment-7fc7848b56-sbkhv   1/1     Running   0          1m
  7. Verify that the Grafana data sources installed correctly:

    $ oc get grafanadatasources
    
    NAME                    AGE
    default-datasources     20h
  8. Verify that the Grafana route exists:

    $ oc get route grafana-route
    
    NAME            HOST/PORT                                          PATH   SERVICES          PORT   TERMINATION   WILDCARD
    grafana-route   grafana-route-service-telemetry.apps.infra.watch          grafana-service   3000   edge          None

Overriding the default Grafana container image

The dashboards in Service Telemetry Framework (STF) require features that are available only in Grafana version 8.1.0 and later. By default, the Service Telemetry Operator installs a compatible version. You can override the base Grafana image by specifying the image path to an image registry with graphing.grafana.baseImage.

Procedure
  1. Ensure that you have the correct version of Grafana:

    $ oc get pod -l "app=grafana" -ojsonpath='{.items[0].spec.containers[0].image}'
    docker.io/grafana/grafana:7.3.10
  2. If the running image is older than 8.1.0, patch the ServiceTelemetry object to update the image. Service Telemetry Operator updates the Grafana manifest, which restarts the Grafana deployment:

    $ oc patch stf/default --type merge -p '{"spec":{"graphing":{"grafana":{"baseImage":"docker.io/grafana/grafana:8.1.5"}}}}'
  3. Verify that a new Grafana pod exists and has a STATUS value of Running:

    $ oc get pod -l "app=grafana"
    NAME                                 READY     STATUS    RESTARTS   AGE
    grafana-deployment-fb9799b58-j2hj2   1/1       Running   0          10s
  4. Verify that the new instance is running the updated image:

    $ oc get pod -l "app=grafana" -ojsonpath='{.items[0].spec.containers[0].image}'
    docker.io/grafana/grafana:8.1.0

Importing dashboards

The Grafana Operator can import and manage dashboards by creating GrafanaDashboard objects. You can view example dashboards at https://github.com/infrawatch/dashboards.

Procedure
  1. Import the infrastructure dashboard:

    $ oc apply -f https://raw.githubusercontent.com/infrawatch/dashboards/master/deploy/stf-1.3/rhos-dashboard.yaml
    
    grafanadashboard.integreatly.org/rhos-dashboard-1.3 created
  2. Import the cloud dashboard:

    For some panels in the cloud dashboard, you must set the value of the collectd virt plugin parameter hostname_format to name uuid hostname in the stf-connectors.yaml file. If you do not configure this parameter, affected dashboards remain empty. For more information about the virt plugin, see collectd plugins.
    $ oc apply -f https://raw.githubusercontent.com/infrawatch/dashboards/master/deploy/stf-1.3/rhos-cloud-dashboard.yaml
    
    grafanadashboard.integreatly.org/rhos-cloud-dashboard-1.3 created
  3. Import the cloud events dashboard:

    $ oc apply -f https://raw.githubusercontent.com/infrawatch/dashboards/master/deploy/stf-1.3/rhos-cloudevents-dashboard.yaml
    
    grafanadashboard.integreatly.org/rhos-cloudevents-dashboard created
  4. Import the virtual machine dashboard:

    $ oc apply -f https://raw.githubusercontent.com/infrawatch/dashboards/master/deploy/stf-1.3/virtual-machine-view.yaml
    
    grafanadashboard.integreatly.org/virtual-machine-view-1.3 configured
  5. Import the memcached dashboard:

    $ oc apply -f https://raw.githubusercontent.com/infrawatch/dashboards/master/deploy/stf-1.3/memcached-dashboard.yaml
    
    grafanadashboard.integreatly.org/memcached-dashboard-1.3 created
  6. Verify that the dashboards are available:

    $ oc get grafanadashboards
    
    NAME                   AGE
    memcached-dashboard-1.3      115s
    rhos-cloud-dashboard-1.3     2m12s
    rhos-cloudevents-dashboard   2m6s
    rhos-dashboard-1.3           2m17s
    virtual-machine-view-1.3     2m
  7. Retrieve the Grafana route address:

    $ oc get route grafana-route -ojsonpath='{.spec.host}'
    
    grafana-route-service-telemetry.apps.infra.watch
  8. In a web browser, navigate to https://<grafana_route_address>. Replace <grafana_route_address> with the value that you retrieved in the previous step.

  9. To view the dashboard, click Dashboards and Manage.

Retrieving and setting Grafana login credentials

Service Telemetry Framework (STF) sets default login credentials when Grafana is enabled. You can override the credentials in the ServiceTelemetry object.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Retrieve the default username and password from the STF object:

    $ oc get stf default -o jsonpath="{.spec.graphing.grafana['adminUser','adminPassword']}"
  4. To modify the default values of the Grafana administrator username and password through the ServiceTelemetry object, use the graphing.grafana.adminUser and graphing.grafana.adminPassword parameters.

Metrics retention time period in Service Telemetry Framework

The default retention time for metrics stored in Service Telemetry Framework (STF) is 24 hours, which provides enough data for trends to develop for the purposes of alerting.

For long-term storage, use systems designed for long-term data retention, for example, Thanos.

Additional resources

Editing the metrics retention time period in Service Telemetry Framework

You can adjust Service Telemetry Framework (STF) for additional metrics retention time.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Edit the ServiceTelemetry object:

    $ oc edit stf default
  4. Add retention: 7d to the storage section of backends.metrics.prometheus.storage to increase the retention period to seven days:

    If you set a long retention period, retrieving data from heavily populated Prometheus systems can result in queries returning results slowly.
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: stf-default
      namespace: service-telemetry
    spec:
      ...
      backends:
        metrics:
          prometheus:
            enabled: true
            storage:
              strategy: persistent
              retention: 7d
        ...
  5. Save your changes and close the object.

Additional resources

Alerts in Service Telemetry Framework

You create alert rules in Prometheus and alert routes in Alertmanager. Alert rules in Prometheus servers send alerts to an Alertmanager, which manages the alerts. Alertmanager can silence, inhibit, or aggregate alerts, and send notifications by using email, on-call notification systems, or chat platforms.

To create an alert, complete the following tasks:

  1. Create an alert rule in Prometheus. For more information, see Creating an alert rule in Prometheus.

  2. Create an alert route in Alertmanager. There are two ways in which you can create an alert route:

Additional resources

For more information about alerts or notifications with Prometheus and Alertmanager, see https://prometheus.io/docs/alerting/overview/

To view an example set of alerts that you can use with Service Telemetry Framework (STF), see https://github.com/infrawatch/service-telemetry-operator/tree/master/deploy/alerts

Creating an alert rule in Prometheus

Prometheus evaluates alert rules to trigger notifications. If the rule condition returns an empty result set, the condition is false. Otherwise, the rule is true and it triggers an alert.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Create a PrometheusRule object that contains the alert rule. The Prometheus Operator loads the rule into Prometheus:

    $ oc apply -f - <<EOF
    apiVersion: monitoring.coreos.com/v1
    kind: PrometheusRule
    metadata:
      creationTimestamp: null
      labels:
        prometheus: default
        role: alert-rules
      name: prometheus-alarm-rules
      namespace: service-telemetry
    spec:
      groups:
        - name: ./openstack.rules
          rules:
            - alert: Collectd metrics receive rate is zero
              expr: rate(sg_total_collectd_msg_received_count[1m]) == 0 (1)
    EOF
    1 To change the rule, edit the value of the expr parameter.
  4. To verify that the Operator loaded the rules into Prometheus, run the curl command against the default-prometheus-proxy route with basic authentication:

    $ curl -k --user "internal:$(oc get secret default-prometheus-htpasswd -ogo-template='{{ .data.password | base64decode }}')" https://$(oc get route default-prometheus-proxy -ogo-template='{{ .spec.host }}')/api/v1/rules
    
    {"status":"success","data":{"groups":[{"name":"./openstack.rules","file":"/etc/prometheus/rules/prometheus-default-rulefiles-0/service-telemetry-prometheus-alarm-rules.yaml","rules":[{"state":"inactive","name":"Collectd metrics receive count is zero","query":"rate(sg_total_collectd_msg_received_count[1m]) == 0","duration":0,"labels":{},"annotations":{},"alerts":[],"health":"ok","evaluationTime":0.00034627,"lastEvaluation":"2021-12-07T17:23:22.160448028Z","type":"alerting"}],"interval":30,"evaluationTime":0.000353787,"lastEvaluation":"2021-12-07T17:23:22.160444017Z"}]}}

Configuring custom alerts

You can add custom alerts to the PrometheusRule object that you created in Creating an alert rule in Prometheus.

Procedure
  1. Use the oc edit command:

    $ oc edit prometheusrules prometheus-alarm-rules
  2. Edit the PrometheusRules manifest.

  3. Save and close the manifest.

Additional resources

Creating a standard alert route in Alertmanager

Use Alertmanager to deliver alerts to an external system, such as email, IRC, or other notification channel. The Prometheus Operator manages the Alertmanager configuration as a OpenShift secret. By default, Service Telemetry Framework (STF) deploys a basic configuration that results in no receivers:

alertmanager.yaml: |-
  global:
    resolve_timeout: 5m
  route:
    group_by: ['job']
    group_wait: 30s
    group_interval: 5m
    repeat_interval: 12h
    receiver: 'null'
  receivers:
  - name: 'null'

To deploy a custom Alertmanager route with STF, you must pass an alertmanagerConfigManifest parameter to the Service Telemetry Operator that results in an updated secret, managed by the Prometheus Operator.

If your alertmanagerConfigManifest contains a custom template to construct the title and text of the sent alert, deploy the contents of the alertmanagerConfigManifest using a base64-encoded configuration. For more information, see Creating an alert route with templating in Alertmanager.
Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Edit the ServiceTelemetry object for your STF deployment:

    $ oc edit stf default
  4. Add the new parameter alertmanagerConfigManifest and the Secret object contents to define the alertmanager.yaml configuration for Alertmanager:

    This step loads the default template that the Service Telemetry Operator manages. To verify that the changes are populating correctly, change a value, return the alertmanager-default secret, and verify that the new value is loaded into memory. For example, change the value of the parameter global.resolve_timeout from 5m to 10m.
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        metrics:
          prometheus:
            enabled: true
      alertmanagerConfigManifest: |
        apiVersion: v1
        kind: Secret
        metadata:
          name: 'alertmanager-default'
          namespace: 'service-telemetry'
        type: Opaque
        stringData:
          alertmanager.yaml: |-
            global:
              resolve_timeout: 10m
            route:
              group_by: ['job']
              group_wait: 30s
              group_interval: 5m
              repeat_interval: 12h
              receiver: 'null'
            receivers:
            - name: 'null'
  5. Verify that the configuration has been applied to the secret:

    $ oc get secret alertmanager-default -o go-template='{{index .data "alertmanager.yaml" | base64decode }}'
    
    global:
      resolve_timeout: 10m
    route:
      group_by: ['job']
      group_wait: 30s
      group_interval: 5m
      repeat_interval: 12h
      receiver: 'null'
    receivers:
    - name: 'null'
  6. Run the curl command against the alertmanager-proxy service to retrieve the status and configYAML contents, and verify that the supplied configuration matches the configuration in Alertmanager:

    $ oc run curl -it --serviceaccount=prometheus-k8s --restart='Never' --image=radial/busyboxplus:curl -- sh -c "curl -k -H \"Content-Type: application/json\" -H \"Authorization: Bearer \$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)\" https://default-alertmanager-proxy:9095/api/v1/status"
    
    {"status":"success","data":{"configYAML":"...",...}}
  7. Verify that the configYAML field contains the changes you expect.

  8. To clean up the environment, delete the curl pod:

    $ oc delete pod curl
    
    pod "curl" deleted
Additional resources

Creating an alert route with templating in Alertmanager

Use Alertmanager to deliver alerts to an external system, such as email, IRC, or other notification channel. The Prometheus Operator manages the Alertmanager configuration as a OpenShift secret. By default, Service Telemetry Framework (STF) deploys a basic configuration that results in no receivers:

alertmanager.yaml: |-
  global:
    resolve_timeout: 5m
  route:
    group_by: ['job']
    group_wait: 30s
    group_interval: 5m
    repeat_interval: 12h
    receiver: 'null'
  receivers:
  - name: 'null'

If the alertmanagerConfigManifest parameter contains a custom template, for example, to construct the title and text of the sent alert, deploy the contents of the alertmanagerConfigManifest by using a base64-encoded configuration.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Edit the ServiceTelemetry object for your STF deployment:

    $ oc edit stf default
  4. To deploy a custom Alertmanager route with STF, you must pass an alertmanagerConfigManifest parameter to the Service Telemetry Operator that results in an updated secret that is managed by the Prometheus Operator:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      backends:
        metrics:
          prometheus:
            enabled: true
      alertmanagerConfigManifest: |
        apiVersion: v1
        kind: Secret
        metadata:
          name: 'alertmanager-default'
          namespace: 'service-telemetry'
        type: Opaque
        data:
          alertmanager.yaml: Z2xvYmFsOgogIHJlc29sdmVfdGltZW91dDogMTBtCiAgc2xhY2tfYXBpX3VybDogPHNsYWNrX2FwaV91cmw+CnJlY2VpdmVyczoKICAtIG5hbWU6IHNsYWNrCiAgICBzbGFja19jb25maWdzOgogICAgLSBjaGFubmVsOiAjc3RmLWFsZXJ0cwogICAgICB0aXRsZTogfC0KICAgICAgICAuLi4KICAgICAgdGV4dDogPi0KICAgICAgICAuLi4Kcm91dGU6CiAgZ3JvdXBfYnk6IFsnam9iJ10KICBncm91cF93YWl0OiAzMHMKICBncm91cF9pbnRlcnZhbDogNW0KICByZXBlYXRfaW50ZXJ2YWw6IDEyaAogIHJlY2VpdmVyOiAnc2xhY2snCg==
  5. Verify that the configuration has been applied to the secret:

    $ oc get secret alertmanager-default -o go-template='{{index .data "alertmanager.yaml" | base64decode }}'
    
    global:
      resolve_timeout: 10m
      slack_api_url: <slack_api_url>
    receivers:
      - name: slack
        slack_configs:
        - channel: #stf-alerts
          title: |-
            ...
          text: >-
            ...
    route:
      group_by: ['job']
      group_wait: 30s
      group_interval: 5m
      repeat_interval: 12h
      receiver: 'slack'
  6. Run the curl command against the alertmanager-proxy service to retrieve the status and configYAML contents, and verify that the supplied configuration matches the configuration in Alertmanager:

    $ oc run curl -it --serviceaccount=prometheus-k8s --restart='Never' --image=radial/busyboxplus:curl -- sh -c "curl -k -H \"Content-Type: application/json\" -H \"Authorization: Bearer \$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)\" https://default-alertmanager-proxy:9095/api/v1/status"
    
    {"status":"success","data":{"configYAML":"...",...}}
  7. Verify that the configYAML field contains the changes you expect.

  8. To clean up the environment, delete the curl pod:

    $ oc delete pod curl
    
    pod "curl" deleted
Additional resources

Configuring SNMP traps

You can integrate Service Telemetry Framework (STF) with an existing infrastructure monitoring platform that receives notifications through SNMP traps. To enable SNMP traps, modify the ServiceTelemetry object and configure the snmpTraps parameters.

For more information about configuring alerts, see Alerts in Service Telemetry Framework.

Prerequisites
  • Know the IP address or hostname of the SNMP trap receiver where you want to send the alerts

Procedure
  1. To enable SNMP traps, modify the ServiceTelemetry object:

    $ oc edit stf default
  2. Set the alerting.alertmanager.receivers.snmpTraps parameters:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    ...
    spec:
      ...
      alerting:
        alertmanager:
          receivers:
            snmpTraps:
              enabled: true
              target: 10.10.10.10
  3. Ensure that you set the value of target to the IP address or hostname of the SNMP trap receiver.

High availability

With high availability, Service Telemetry Framework (STF) can rapidly recover from failures in its component services. Although OpenShift restarts a failed pod if nodes are available to schedule the workload, this recovery process might take more than one minute, during which time events and metrics are lost. A high availability configuration includes multiple copies of STF components, which reduces recovery time to approximately 2 seconds. To protect against failure of an OpenShift node, deploy STF to an OpenShift cluster with three or more nodes.

STF is not yet a fully fault tolerant system. Delivery of metrics and events during the recovery period is not guaranteed.

Enabling high availability has the following effects:

  • Three ElasticSearch pods run instead of the default one.

  • The following components run two pods instead of the default one:

    • Apache Qpid Dispatch Router

    • Alertmanager

    • Prometheus

    • Events Smart Gateway

    • Metrics Smart Gateway

  • Recovery time from a lost pod in any of these services reduces to approximately 2 seconds.

Configuring high availability

To configure Service Telemetry Framework (STF) for high availability, add highAvailability.enabled: true to the ServiceTelemetry object in OpenShift. You can set this parameter at installation time or, if you already deployed STF, complete the following steps:

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Use the oc command to edit the ServiceTelemetry object:

    $ oc edit stf default
  4. Add highAvailability.enabled: true to the spec section:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    ...
    spec:
      ...
      highAvailability:
        enabled: true
  5. Save your changes and close the object.

Ephemeral storage

You can use ephemeral storage to run Service Telemetry Framework (STF) without persistently storing data in your OpenShift cluster.

If you use ephemeral storage, you might experience data loss if a pod is restarted, updated, or rescheduled onto another node. Use ephemeral storage only for development or testing, and not production environments.

Configuring ephemeral storage

To configure STF components for ephemeral storage, add ...storage.strategy: ephemeral to the corresponding parameter. For example, to enable ephemeral storage for the Prometheus back end, set backends.metrics.prometheus.storage.strategy: ephemeral. Components that support configuration of ephemeral storage include alerting.alertmanager, backends.metrics.prometheus, and backends.events.elasticsearch. You can add ephemeral storage configuration at installation time or, if you already deployed STF, complete the following steps:

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Edit the ServiceTelemetry object:

    $ oc edit stf default
  4. Add the ...storage.strategy: ephemeral parameter to the spec section of the relevant component:

    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: stf-default
      namespace: service-telemetry
    spec:
      alerting:
        enabled: true
        alertmanager:
          storage:
            strategy: ephemeral
      backends:
        metrics:
          prometheus:
            enabled: true
            storage:
              strategy: ephemeral
        events:
          elasticsearch:
            enabled: true
            storage:
              strategy: ephemeral
  5. Save your changes and close the object.

Observability Strategy in Service Telemetry Framework

Service Telemetry Framework (STF) does not include storage backends and alerting tools. STF uses community operators to deploy Prometheus, Alertmanager, Grafana, and Elasticsearch. STF makes requests to these community operators to create instances of each application configured to work with STF.

Instead of having Service Telemetry Operator create custom resource requests, you can use your own deployments of these applications or other compatible applications, and scrape the metrics Smart Gateways for delivery to your own Prometheus-compatible system for telemetry storage. If you set the observability strategy to use alternative backends instead, persistent or ephemeral storage is not required for STF.

Configuring an alternate observability strategy

To configure STF to skip the deployment of storage, visualization, and alerting backends, add observabilityStrategy: none to the ServiceTelemetry spec. In this mode, only Apache Qpid Dispatch Router routers and metrics Smart Gateways are deployed, and you must configure an external Prometheus-compatible system to collect metrics from the STF Smart Gateways.

Currently, only metrics are supported when you set observabilityStrategy to none. Events Smart Gateways are not deployed.
Procedure
  1. Create a ServiceTelemetry object with the property observabilityStrategy: none in the spec parameter. The manifest shows results in a default deployment of STF that is suitable for receiving telemetry from a single cloud with all metrics collector types.

    $ oc apply -f - <<EOF
    apiVersion: infra.watch/v1beta1
    kind: ServiceTelemetry
    metadata:
      name: default
      namespace: service-telemetry
    spec:
      observabilityStrategy: none
    EOF
  2. To verify that all workloads are operating correctly, view the pods and the status of each pod:

    $ oc get pods
    NAME                                                      READY   STATUS    RESTARTS   AGE
    default-cloud1-ceil-meter-smartgateway-59c845d65b-gzhcs   3/3     Running   0          132m
    default-cloud1-coll-meter-smartgateway-75bbd948b9-d5phm   3/3     Running   0          132m
    default-cloud1-sens-meter-smartgateway-7fdbb57b6d-dh2g9   3/3     Running   0          132m
    default-interconnect-668d5bbcd6-57b2l                     1/1     Running   0          132m
    interconnect-operator-b8f5bb647-tlp5t                     1/1     Running   0          47h
    service-telemetry-operator-566b9dd695-wkvjq               1/1     Running   0          156m
    smart-gateway-operator-58d77dcf7-6xsq7                    1/1     Running   0          47h
Additional resources

For more information about configuring additional clouds or to change the set of supported collectors, see Deploying Smart Gateways

Configuring openshift-monitoring to consume metrics from STF

You can configure openshift-monitoring to consume metrics from STF so that you can use the existing Prometheus deployment for STF data. This configuration is useful in combination with observabilityStrategy: none as an alternative to the community operators. You must add a label to the namespace where STF is deployed, and create ServiceMonitor objects for each Smart Gateway intended to be scraped.

Procedure
  1. Edit the namespace object:

    $ oc edit namespace service-telemetry
  2. Add the openshift.io/cluster-monitoring: "true" label under the metadata property:

    metadata:
      labels:
        openshift.io/cluster-monitoring: "true"
  3. Create a ServiceMonitor object for each Smart Gateway:

    $ for collector_type in ceil coll sens; do oc apply -f <(sed -e "s/<<COLLECTOR_TYPE>>/${collector_type}/g" << EOF
    apiVersion: monitoring.coreos.com/v1
    kind: ServiceMonitor
    metadata:
      labels:
        app: smart-gateway
      name: default-cloud1-<<COLLECTOR_TYPE>>-meter
      namespace: service-telemetry
    spec:
      endpoints:
      - bearerTokenFile: /var/run/secrets/kubernetes.io/serviceaccount/token
        interval: 10s
        metricRelabelings:
        - action: labeldrop
          regex: pod
        - action: labeldrop
          regex: namespace
        - action: labeldrop
          regex: instance
        - action: labeldrop
          regex: job
        - action: labeldrop
          regex: publisher
        port: prom-https
        scheme: https
        tlsConfig:
          caFile: /etc/prometheus/configmaps/serving-certs-ca-bundle/service-ca.crt
          serverName: default-cloud1-<<COLLECTOR_TYPE>>-meter.service-telemetry.svc
      namespaceSelector:
        matchNames:
        - service-telemetry
      selector:
        matchLabels:
          app: smart-gateway
          smart-gateway: default-cloud1-<<COLLECTOR_TYPE>>-meter
    EOF
    ); done
    servicemonitor.monitoring.coreos.com/default-cloud1-ceil-meter configured
    servicemonitor.monitoring.coreos.com/default-cloud1-coll-meter configured
    servicemonitor.monitoring.coreos.com/default-cloud1-sens-meter configured
  4. To verify the successful configuration of openshift-monitoring, ensure that Smart Gateway metrics appear in Prometheus.

  5. Retrieve the route for the openshift-monitoring prometheus:

    $ oc get route -n openshift-monitoring prometheus-k8s
    NAME             HOST/PORT                                              PATH   SERVICES         PORT   TERMINATION          WILDCARD
    prometheus-k8s   prometheus-k8s-openshift-monitoring.apps.infra.watch          prometheus-k8s   web    reencrypt/Redirect   None
  6. Visit the host in your browser and log in with OpenShift credentials.

  7. Verify that the following targets are visible under the Status -> Targets tab:

    • service-telemetry/default-cloud1-ceil-meter/0

    • service-telemetry/default-cloud1-coll-meter/0

    • service-telemetry/default-cloud1-sens-meter/0

      If there are problems with the configuration, find them on this page.

  8. Issue the following queries on the Graph tab:

    • sg_total_collectd_metric_decode_count

    • sg_total_ceilometer_metric_decode_count

    • sg_total_sensubility_metric_decode_count

  9. There should be one result from each Smart Gateway, as shown in the following example:

    If the values returned are 0, it means that STF is not receiving that type of metric yet but as long as a result is returned, the configuration of openshift-monitoring is correct.
    • sg_total_collectd_metric_decode_count{container="sg-core", endpoint="prom-https", service="default-cloud1-coll-meter", source="SG"}

    • sg_total_ceilometer_metric_decode_count{container="sg-core", endpoint="prom-https", service="default-cloud1-ceil-meter", source="SG"}

    • sg_total_sensubility_metric_decode_count{container="sg-core", endpoint="prom-https", service="default-cloud1-sens-meter", source="SG"}

Resource usage of OpenStack services

You can monitor the resource usage of the OpenStack (OSP) services, such as the APIs and other infrastructure processes, to identify bottlenecks in the overcloud by showing services that run out of compute power. Resource usage monitoring is enabled by default.

Additional resources

Disabling resource usage monitoring of OpenStack services

To disable the monitoring of OSP containerized service resource usage, you must set the CollectdEnableLibpodstats parameter to false.

Prerequisites
Procedure
  1. Open the stf-connectors.yaml file and add the CollectdEnableLibpodstats parameter to override the setting in enable-stf.yaml. Ensure that stf-connectors.yaml is called from the openstack overcloud deploy command after enable-stf.yaml:

      CollectdEnableLibpodstats: false
  2. Continue with the overcloud deployment procedure. For more information, see Deploying the overcloud.

OpenStack API status and containerized services health

You can use the OCI (Open Container Initiative) standard to assess the container health status of each OpenStack (OSP) service by periodically running a health check script. Most OSP services implement a health check that logs issues and returns a binary status. For the OSP APIs, the health checks query the root endpoint and determine the health based on the response time.

Monitoring of OSP container health and API status is enabled by default.

Additional resources

Disabling container health and API status monitoring

To disable OSP containerized service health and API status monitoring, you must set the CollectdEnableSensubility parameter to false.

Prerequisites
Procedure
  1. Open the stf-connectors.yaml and add the CollectdEnableSensubility parameter to override the setting in enable-stf.yaml. Ensure that stf-connectors.yaml is called from the openstack overcloud deploy command after enable-stf.yaml:

    CollectdEnableSensubility: false
  2. Continue with the overcloud deployment procedure. For more information, see Deploying the overcloud.

Additional resources

Upgrading Service Telemetry Framework to version 1.4

To migrate from Service Telemetry Framework (STF) v1.3 to STF v1.4, you must replace the ClusterServiceVersion and Subscription objects in the service-telemetry namespace on your OpenShift environment.

Prerequisites
  • You have upgraded your OpenShift environment to v4.8. STF v1.4 does not run on OpenShift versions less than v4.8.

  • You have backed up your data. Upgrading STF v1.3 to v1.4 results in a brief outage while the Smart Gateways and other components are updated. Additionally, changes to the ServiceTelemetry and SmartGateway objects do not have any effect while the Operators are being replaced.

To upgrade from STF v1.3 to v1.4, complete the following procedures:

Removing STF 1.3 Smart Gateway Operator

Remove the Smart Gateway Operator from STF 1.3.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Retrieve the Subscription name of the Smart Gateway Operator. Replace service-telemetry in the selector with the namespace that hosts your STF instance if it is different from the default namespace. Verify that only one subscription is returned:

    $ oc get sub --selector=operators.coreos.com/smart-gateway-operator.service-telemetry
    
    NAME                                                                       PACKAGE                  SOURCE             CHANNEL
    smart-gateway-operator-stable-1.3-redhat-operators-openshift-marketplace   smart-gateway-operator   redhat-operators   stable-1.3
  4. Delete the Smart Gateway Operator subscription:

    $ oc delete sub --selector=operators.coreos.com/smart-gateway-operator.service-telemetry
    
    subscription.operators.coreos.com "smart-gateway-operator-stable-1.3-redhat-operators-openshift-marketplace" deleted
  5. Retrieve the Smart Gateway Operator ClusterServiceVersion and verify that only one ClusterServiceVersion is returned:

    $ oc get csv --selector=operators.coreos.com/smart-gateway-operator.service-telemetry
    
    NAME                                     DISPLAY                  VERSION          REPLACES   PHASE
    smart-gateway-operator.v3.0.1635451893   Smart Gateway Operator   3.0.1635451893              Succeeded
  6. Delete the Smart Gateway Operator ClusterServiceVersion:

    $ oc delete csv --selector=operators.coreos.com/smart-gateway-operator.service-telemetry
    
    clusterserviceversion.operators.coreos.com "smart-gateway-operator.v3.0.1635451893" deleted
  7. Delete the SmartGateway Custom Resource Definition (CRD). After removal of the CRD, no data flows to STF until the upgrade is completed and the Smart Gateway instances are reinstantiated:

    $ oc delete crd smartgateways.smartgateway.infra.watch
    
    customresourcedefinition.apiextensions.k8s.io "smartgateways.smartgateway.infra.watch" deleted

Updating the Service Telemetry Operator to 1.4

You must change the subscription channel of the Service Telemetry Operator which manages the STF instances to the stable-1.4 channel.

Procedure
  1. Log in to OpenShift.

  2. Change to the service-telemetry namespace:

    $ oc project service-telemetry
  3. Patch the Service Telemetry Operator Subscription to use the stable-1.4 channel. Replace the service-telemetry in the selector with the namespace that hosts your STF instance if it is different from the default namespace:

    $ oc patch $(oc get sub --selector=operators.coreos.com/service-telemetry-operator.service-telemetry -oname) --patch $'spec:\n  channel: stable-1.4' --type=merge
    
    subscription.operators.coreos.com/service-telemetry-operator patched
  4. Monitor the output of the oc get csv command until the Smart Gateway Operator is installed and Service Telemetry Operator moves through the update phases. When the phase changes to Succeeded, the Service Telemetry Operator has completed the update:

    $ watch -n5 oc get csv
    
    NAME                                         DISPLAY                                         VERSION          REPLACES                                     PHASE
    amq7-cert-manager.v1.0.3                     Red Hat Integration - AMQ Certificate Manager   1.0.3            amq7-cert-manager.v1.0.2                     Succeeded
    amq7-interconnect-operator.v1.10.5           Red Hat Integration - AMQ Interconnect          1.10.5           amq7-interconnect-operator.v1.10.4           Succeeded
    elasticsearch-eck-operator-certified.1.9.1   Elasticsearch (ECK) Operator                    1.9.1                                                         Succeeded
    prometheusoperator.0.47.0                    Prometheus Operator                             0.47.0           prometheusoperator.0.37.0                    Succeeded
    service-telemetry-operator.v1.4.1641504218   Service Telemetry Operator                      1.4.1641504218   service-telemetry-operator.v1.3.1635451892   Succeeded
    smart-gateway-operator.v4.0.1641504216       Smart Gateway Operator                          4.0.1641504216                                                Succeeded
  5. Validate that all pods are ready and running. Your environment might differ from the following example output:

    $ oc get pods
    
    NAME                                                      READY   STATUS    RESTARTS   AGE
    alertmanager-default-0                                    3/3     Running   0          162m
    default-cloud1-ceil-event-smartgateway-5599bcfc9d-wp48n   2/2     Running   1          160m
    default-cloud1-ceil-meter-smartgateway-c8fdf579c-955kt    3/3     Running   0          160m
    default-cloud1-coll-event-smartgateway-97b54b7dc-5zz2v    2/2     Running   0          159m
    default-cloud1-coll-meter-smartgateway-774b9988b8-wb5vd   3/3     Running   0          160m
    default-cloud1-sens-meter-smartgateway-b98966fbf-rnqwf    3/3     Running   0          159m
    default-interconnect-675dd97bc4-dcrzk                     1/1     Running   0          171m
    default-snmp-webhook-7854d4889d-wgmgg                     1/1     Running   0          171m
    elastic-operator-c54ff8cc-jcg8d                           1/1     Running   6          3h55m
    elasticsearch-es-default-0                                1/1     Running   0          160m
    interconnect-operator-6bf74c4ffb-hkmbq                    1/1     Running   0          3h54m
    prometheus-default-0                                      3/3     Running   1          160m
    prometheus-operator-fc64987d-f7gx4                        1/1     Running   0          3h54m
    service-telemetry-operator-68d888f767-s5kzh               1/1     Running   0          163m
    smart-gateway-operator-584df7959-llxgl                    1/1     Running   0          163m

collectd plugins

You can configure multiple collectd plugins depending on your OpenStack (OSP) Train environment.

The following list of plugins shows the available heat template ExtraConfig parameters that you can set to override the defaults. Each section provides the general configuration name for the ExtraConfig option. For example, if there is a collectd plugin called example_plugin, the format of the plugin title is collectd::plugin::example_plugin.

Reference the tables of available parameters for specific plugins, such as in the following example:

ExtraConfig:
  collectd::plugin::example_plugin::<parameter>: <value>

Reference the metrics tables of specific plugins for Prometheus or Grafana queries.

collectd::plugin::aggregation

You can aggregate several values into one with the aggregation plugin. Use the aggregation functions such as sum, average, min, and max to calculate metrics, for example average and total CPU statistics.

Table 2. aggregation parameters
Parameter Type

host

String

plugin

String

plugininstance

Integer

agg_type

String

typeinstance

String

sethost

String

setplugin

String

setplugininstance

Integer

settypeinstance

String

groupby

Array of Strings

calculatesum

Boolean

calculatenum

Boolean

calculateaverage

Boolean

calculateminimum

Boolean

calculatemaximum

Boolean

calculatestddev

Boolean

Example configuration:

Deploy three aggregate configurations which results in generation of files:

  1. aggregator-calcCpuLoadAvg.conf

  2. aggregator-calcCpuLoadMinMax.conf

  3. aggregator-calcMemoryTotalMaxAvg.conf

The aggregation configurations use the default CPU and Memory plugin configurations. The following aggregations are created:

  1. Calculate average CPU load for all CPU cores grouped by host and state.

  2. Calculate minimum and maxiumum CPU load groups by host and state.

  3. Calculate maximum, average, and total for memory grouped by type.

parameter_defaults:
  CollectdExtraPlugins:
    - aggregation

  ExtraConfig:
    collectd::plugin::aggregation::aggregators:
      calcCpuLoadAvg:
        plugin: "cpu"
        agg_type: "cpu"
        groupby:
          - "Host"
          - "TypeInstance"
        calculateaverage: True
      calcCpuLoadMinMax:
        plugin: "cpu"
        agg_type: "cpu"
        groupby:
          - "Host"
          - "TypeInstance"
        calculatemaximum: True
        calculateminimum: True
      calcMemoryTotalMaxAvg:
        plugin: "memory"
        agg_type: "memory"
        groupby:
          - "TypeInstance"
        calculatemaximum: True
        calculateaverage: True
        calculatesum: True

collectd::plugin::ampq

collectd::plugin::amqp1

Use the amqp1 plugin to write values to an amqp1 message bus, for example, Apache Qpid Dispatch Router.

Table 3. amqp1 parameters
Parameter Type

manage_package

Boolean

transport

String

host

string

port

integer

user

String

password

String

address

String

instances

Hash

retry_delay

Integer

send_queue_limit

Integer

interval

Integer

Use the send_queue_limit parameter to limit the length of the outgoing metrics queue.

If there is no AMQP1 connection, the plugin continues to queue messages to send, which can result in unbounded memory consumption. The default value is 0, which disables the outgoing metrics queue.

Increase the value of the send_queue_limit parameter if metrics are missing.

Example configuration:
  Parameter_defaults:
    CollectdExtraPlugins:
      - amqp1
    ExtraConfig:
      collectd::plugin::amqp1::send_queue_limit: 5000

collectd::plugin::apache

Use the apache plugin to collect Apache data.

Table 4. apache parameters
Parameter Type

instances

Hash

interval

Integer

manage-package

Boolean

package_install_options

List

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::apache:
          localhost:
              url: "http://10.0.0.111/status?auto"
Additional resources

For more information about configuring the apache plugin, see apache.

collectd::plugin::battery

Use the battery plugin to report the remaining capacity, power, or voltage of laptop batteries.

Table 5. battery parameters
Parameter Type

values_percentage

Boolean

report_degraded

Boolean

query_state_fs

Boolean

interval

Integer

Additional resources

For more information about configuring the battery plugin, see battery.

collectd::plugin::bind

Use the bind plugin to retrieve encoded statistics about queries and responses from a DNS server. The plugin submits the values to collectd.

collectd::plugin::ceph

Use the ceph plugin to gather data from ceph daemons.

Table 6. ceph parameters
Parameter Type

daemons

Array

longrunavglatency

Boolean

convertspecialmetrictypes

Boolean

manage_package

Boolean

package_name

String

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::ceph::daemons:
           - ceph-osd.0
           - ceph-osd.1
           - ceph-osd.2
           - ceph-osd.3
           - ceph-osd.4
If an Object Storage Daemon (OSD) is not on every node, you must list the OSDs.
When you deploy collectd, the ceph plugin is added to the Ceph nodes. Do not add the ceph plugin on Ceph nodes to CollectdExtraPlugins, because this results in a deployment failure.
Additional resources

For more information about configuring the ceph plugin, see ceph.

collectd::plugins::cgroups

Use the cgroups plugin to collect information for processes in a cgroup.

Table 7. cgroups parameters
Parameter Type

ignore_selected

Boolean

interval

Integer

cgroups

List

Additional resources

For more information about configuring the cgroups plugin, see cgroups.

collectd::plugin::connectivity

Use the connectivity plugin to monitor the state of network interfaces.

If no interfaces are listed, all interfaces are monitored by default.
Table 8. connectivity parameters
Parameter Type

interfaces

Array

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::connectivity::interfaces:
        - eth0
        - eth1
Additional resources

For more information about configuring the connectivity plugin, see connectivity.

collectd::plugin::conntrack

Use the conntrack plugin to track the number of entries in the Linux connection-tracking table. There are no parameters for this plugin.

collectd::plugin::contextswitch

Use the ContextSwitch plugin to collect the number of context switches that the system handles.

Additional resources

For more information about configuring the contextswitch plugin, see contextswitch.

collectd::plugin::cpu

Use the cpu plugin to monitor the time that the CPU spends in various states, for example, idle, executing user code, executing system code, waiting for IO-operations, and other states.

The cpu plugin collects _jiffies_, not percentage values. The value of a jiffy depends on the clock frequency of your hardware platform, and therefore is not an absolute time interval unit.

To report a percentage value, set the Boolean parameters reportbycpu and reportbystate to true, and then set the Boolean parameter valuespercentage to true.

Table 9. cpu metrics
Name Description Query

idle

Amount of idle time

collectd_cpu_total{…​,type_instance=idle}

interrupt

CPU blocked by interrupts

collectd_cpu_total{…​,type_instance=interrupt}

nice

Amount of time running low priority processes

collectd_cpu_total{…​,type_instance=nice}

softirq

Amount of cycles spent in servicing interrupt requests

collectd_cpu_total{…​,type_instance=waitirq}

steal

The percentage of time a virtual CPU waits for a real CPU while the hypervisor is servicing another virtual processor

collectd_cpu_total{…​,type_instance=steal}

system

Amount of time spent on system level (kernel)

collectd_cpu_total{…​,type_instance=system}

user

Jiffies that user processes use

collectd_cpu_total{…​,type_instance=user}

wait

CPU waiting on outstanding I/O request

collectd_cpu_total{…​,type_instance=wait}

Table 10. cpu parameters
Parameter Type

reportbystate

Boolean

valuespercentage

Boolean

reportbycpu

Boolean

reportnumcpu

Boolean

reportgueststate

Boolean

subtractgueststate

Boolean

interval

Integer

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
      - cpu
    ExtraConfig:
        collectd::plugin::cpu::reportbystate: true
Additional resources

For more information about configuring the cpu plugin, see cpu.

collectd::plugin::cpufreq

  • None

collectd::plugin::cpusleep

collectd::plugin::csv

  • collectd::plugin::csv::datadir

  • collectd::plugin::csv::storerates

  • collectd::plugin::csv::interval

collectd::plugin::curl_json

collectd::plugin::curl

collectd::plugin::curl_xml

collectd::plugin::dbi

collectd::plugin::df

Use the df plugin to collect disk space usage information for file systems.

Table 11. df metrics
Name Description Query

free

Amount of free disk space

collectd_df_df_complex{…​, type_instance="free"}

reserved

Amount of reserved disk space

collectd_df_df_complex{…​, type_instance="reserved"}

used

Amount of used disk space

collectd_df_df_complex{…​, type_instance="used"}

Table 12. df parameters
Parameter Type

devices

Array

fstypes

Array

ignoreselected

Boolean

mountpoints

Array

reportbydevice

Boolean

reportinodes

Boolean

reportreserved

Boolean

valuesabsolute

Boolean

valuespercentage

Boolean

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
      - df
    ExtraConfig:
        collectd::plugin::df::fstypes: ['tmpfs','xfs']
Additional resources

For more information about configuring the df plugin, see df.

collectd::plugin::disk

Use the disk plugin to collect performance statistics of hard disks and, if supported, partitions. This plugin is enabled by default.

Table 13. disk parameters
Parameter Type

disks

Array

ignoreselected

Boolean

udevnameattr

String

Table 14. disk metrics
Name Description

merged

The number of operations that can be merged together, already queued operations, for example, one physical disk access served two or more logical operations.

time

The average time an I/O-operation takes to complete. The values might not be fully accurate.

io_time

Time spent doing I/Os (ms). You can use this metric as a device load percentage. A value of 1 second matches 100% of load.

weighted_io_time

Measure of both I/O completion time and the backlog that might be accumulating.

pending_operations

Shows queue size of pending I/O operations.

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::disk::disk: "sda"
        collectd::plugin::disk::ignoreselected: false
Additional resources

For more information about configuring the disk plugin, see disk.

collectd::plugin::dns

collectd::plugin::entropy

  • collectd::plugin::entropy::interval

collectd::plugin::ethstat

  • collectd::plugin::ethstat::interfaces

  • collectd::plugin::ethstat::maps

  • collectd::plugin::ethstat::mappedonly

  • collectd::plugin::ethstat::interval

collectd::plugin::exec

  • collectd::plugin::exec::commands

  • collectd::plugin::exec::commands_defaults

  • collectd::plugin::exec::globals

  • collectd::plugin::exec::interval

collectd::plugin::fhcount

  • collectd::plugin::fhcount::valuesabsolute

  • collectd::plugin::fhcount::valuespercentage

  • collectd::plugin::fhcount::interval

collectd::plugin::filecount

  • collectd::plugin::filecount::directories

  • collectd::plugin::filecount::interval

collectd::plugin::fscache

  • None

collectd-hddtemp

  • collectd::plugin::hddtemp::host

  • collectd::plugin::hddtemp::port

  • collectd::plugin::hddtemp::interval

collectd::plugin::hugepages

Use the hugepages plugin to collect hugepages information. This plugin is enabled by default.

Table 15. hugepages parameters
Parameter Type Defaults

report_per_node_hp

Boolean

true

report_root_hp

Boolean

true

values_pages

Boolean

true

values_bytes

Boolean

false

values_percentage

Boolean

false

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::hugepages::values_percentage: true
Additional resources
  • For more information about configuring the hugepages plugin, see hugepages.

collectd::plugin::intel_rdt

collectd::plugin::interface

Use the interface plugin to measure interface traffic in octets, packets per second, and error rate per second. This plugin is enabled by default.

Table 16. interface parameters
Parameter Type

Default

interfaces

Array

[]

ignoreselected

Boolean

false

reportinactive

Boolean

true

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::interface::interfaces:
           - lo
        collectd::plugin::interface::ignoreselected: true
Additional resources
  • For more information about configuring the interfaces plugin, see interfaces.

collectd::plugin::ipc

  • None

collectd::plugin::ipmi

  • collectd::plugin::ipmi::ignore_selected

  • collectd::plugin::ipmi::notify_sensor_add

  • collectd::plugin::ipmi::notify_sensor_remove

  • collectd::plugin::ipmi::notify_sensor_not_present

  • collectd::plugin::ipmi::sensors

  • collectd::plugin::ipmi::interval

collectd::plugin::iptables

collectd::plugin::irq

  • collectd::plugin::irq::irqs

  • collectd::plugin::irq::ignoreselected

  • collectd::plugin::irq::interval

collectd::plugin::load

Use the load plugin to collect the system load and an overview of the system use. This plugin is enabled by default.

Table 17. plugin parameters
Parameter Type

report_relative

Boolean

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::load::report_relative: false
Additional resources
  • For more information about configuring the load plugin, see load.

collectd::plugin::logfile

  • collectd::plugin::logfile::log_level

  • collectd::plugin::logfile::log_file

  • collectd::plugin::logfile::log_timestamp

  • collectd::plugin::logfile::print_severity

  • collectd::plugin::logfile::interval

collectd::plugin::log_logstash

collectd::plugin::madwifi

collectd::plugin::match_empty_counter

collectd::plugin::match_hashed

collectd::plugin::match_regex

collectd::plugin::match_timediff

collectd::plugin::match_value

collectd::plugin::mbmon

collectd::plugin::mcelog

Use the mcelog plugin to send notifications and statistics that are relevant to Machine Check Exceptions when they occur. Configure mcelog to run in daemon mode and enable logging capabilities.

Table 18. mcelog parameters
Parameter Type

Mcelogfile

String

Memory

Hash { mcelogclientsocket[string], persistentnotification[boolean] }

Example configuration:
parameter_defaults:
    CollectdExtraPlugins: mcelog
    CollectdEnableMcelog: true
Additional resources
  • For more information about configuring the mcelog plugin, see mcelog.

collectd::plugin::md

collectd::plugin::memcachec

collectd::plugin::memcached

  • collectd::plugin::memcached::instances

  • collectd::plugin::memcached::interval

collectd::plugin::memory

The memory plugin provides information about the memory of the system. This plugin is enabled by default.

Table 19. memory parameters
Parameter Type

valuesabsolute

Boolean

valuespercentage

Boolean

Example configuration:
parameter_defaults:
    ExtraConfig:
        collectd::plugin::memory::valuesabsolute: true
        collectd::plugin::memory::valuespercentage: false
Additional resources
  • For more information about configuring the memory plugin, see memory.

collectd::plugin::multimeter

collectd::plugin::mysql

  • collectd::plugin::mysql::interval

  • collectd::plugin::netlink::interfaces

  • collectd::plugin::netlink::verboseinterfaces

  • collectd::plugin::netlink::qdiscs

  • collectd::plugin::netlink::classes

  • collectd::plugin::netlink::filters

  • collectd::plugin::netlink::ignoreselected

  • collectd::plugin::netlink::interval

collectd::plugin::network

  • collectd::plugin::network::timetolive

  • collectd::plugin::network::maxpacketsize

  • collectd::plugin::network::forward

  • collectd::plugin::network::reportstats

  • collectd::plugin::network::listeners

  • collectd::plugin::network::servers

  • collectd::plugin::network::interval

collectd::plugin::nfs

  • collectd::plugin::nfs::interval

collectd::plugin::notify_nagios

collectd::plugin::ntpd

  • collectd::plugin::ntpd::host

  • collectd::plugin::ntpd::port

  • collectd::plugin::ntpd::reverselookups

  • collectd::plugin::ntpd::includeunitid

  • collectd::plugin::ntpd::interval

collectd::plugin::numa

  • None

collectd::plugin::olsrd

collectd::plugin::openldap

collectd::plugin::openvpn

  • collectd::plugin::openvpn::statusfile

  • collectd::plugin::openvpn::improvednamingschema

  • collectd::plugin::openvpn::collectcompression

  • collectd::plugin::openvpn::collectindividualusers

  • collectd::plugin::openvpn::collectusercount

  • collectd::plugin::openvpn::interval

collectd::plugin::ovs_stats

Use the ovs_stats plugin to collect statistics of OVS-connected interfaces. The ovs_stats plugin uses the OVSDB management protocol (RFC7047) monitor mechanism to get statistics from OVSDB.

Table 20. ovs_stats parameters
Parameter Type

address

String

bridges

List

port

Integer

socket

String

Example configuration:

The following example shows how to enable the ovs_stats plugin. If you deploy your overcloud with OVS, you do not need to enable the ovs_stats plugin.

    parameter_defaults:
        CollectdExtraPlugins:
          - ovs_stats
        ExtraConfig:
          collectd::plugin::ovs_stats::socket: '/run/openvswitch/db.sock'
Additional resources
  • For more information about configuring the ovs_stats plugin, see ovs_stats.

collectd::plugin::pcie_errors

Use the pcie_errors plugin to poll PCI config space for baseline and Advanced Error Reporting (AER) errors, and to parse syslog for AER events. Errors are reported through notifications.

Table 21. pcie_errors parameters
Parameter Type

source

Enum (sysfs, proc)

access

String

reportmasked

Boolean

persistent_notifications

Boolean

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
       - pcie_errors
Additional resources
  • For more information about configuring the pcie_errors plugin, see pcie_errors.

collectd::plugin::ping

  • collectd::plugin::ping::hosts

  • collectd::plugin::ping::timeout

  • collectd::plugin::ping::ttl

  • collectd::plugin::ping::source_address

  • collectd::plugin::ping::device

  • collectd::plugin::ping::max_missed

  • collectd::plugin::ping::size

  • collectd::plugin::ping::interval

collectd::plugin::powerdns

  • collectd::plugin::powerdns::interval

  • collectd::plugin::powerdns::servers

  • collectd::plugin::powerdns::recursors

  • collectd::plugin::powerdns::local_socket

  • collectd::plugin::powerdns::interval

collectd::plugin::processes

The processes plugin provides information about system processes. This plugin is enabled by default.

Table 22. plugin parameters
Parameter Type

processes

Array

process_matches

Array

collect_context_switch

Boolean

collect_file_descriptor

Boolean

collect_memory_maps

Boolean

Additional resources
  • For more information about configuring the processes plugin, see processes.

collectd::plugin::protocols

  • collectd::plugin::protocols::ignoreselected

  • collectd::plugin::protocols::values

collectd::plugin::python

collectd::plugin::sensors

collectd::plugin::serial

collectd::plugin::smart

  • collectd::plugin::smart::disks

  • collectd::plugin::smart::ignoreselected

  • collectd::plugin::smart::interval

collectd::plugin::snmp

collectd::plugin::snmp_agent

Use the snmp_agent plugin as an SNMP subagent to map collectd metrics to relevant OIDs. The snmp agent also requires a running snmpd service.

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
        snmp_agent
resource_registry:
    OS::TripleO::Services::Snmp: /usr/share/openstack-tripleo-heat-
templates/deployment/snmp/snmp-baremetal-puppet.yaml
Additional resources:

For more information about how to configure snmp_agent, see snmp_agent.

collectd::plugin::statsd

  • collectd::plugin::statsd::host

  • collectd::plugin::statsd::port

  • collectd::plugin::statsd::deletecounters

  • collectd::plugin::statsd::deletetimers

  • collectd::plugin::statsd::deletegauges

  • collectd::plugin::statsd::deletesets

  • collectd::plugin::statsd::countersum

  • collectd::plugin::statsd::timerpercentile

  • collectd::plugin::statsd::timerlower

  • collectd::plugin::statsd::timerupper

  • collectd::plugin::statsd::timersum

  • collectd::plugin::statsd::timercount

  • collectd::plugin::statsd::interval

collectd::plugin::swap

  • collectd::plugin::swap::reportbydevice

  • collectd::plugin::swap::reportbytes

  • collectd::plugin::swap::valuesabsolute

  • collectd::plugin::swap::valuespercentage

  • collectd::plugin::swap::reportio

  • collectd::plugin::swap::interval

collectd::plugin::sysevent

collectd::plugin::syslog

  • collectd::plugin::syslog::log_level

  • collectd::plugin::syslog::notify_level

  • collectd::plugin::syslog::interval

collectd::plugin::table

  • collectd::plugin::table::tables

  • collectd::plugin::table::interval

collectd::plugin::tail

  • collectd::plugin::tail::files

  • collectd::plugin::tail::interval

collectd::plugin::tail_csv

  • collectd::plugin::tail_csv::metrics

  • collectd::plugin::tail_csv::files

collectd::plugin::target_notification

collectd::plugin::target_replace

collectd::plugin::target_scale

collectd::plugin::target_set

collectd::plugin::target_v5upgrade

collectd::plugin::tcpconns

  • collectd::plugin::tcpconns::localports

  • collectd::plugin::tcpconns::remoteports

  • collectd::plugin::tcpconns::listening

  • collectd::plugin::tcpconns::allportssummary

  • collectd::plugin::tcpconns::interval

collectd::plugin::ted

collectd::plugin::thermal

  • collectd::plugin::thermal::devices

  • collectd::plugin::thermal::ignoreselected

  • collectd::plugin::thermal::interval

collectd::plugin::threshold

  • collectd::plugin::threshold::types

  • collectd::plugin::threshold::plugins

  • collectd::plugin::threshold::hosts

  • collectd::plugin::threshold::interval

collectd::plugin::turbostat

  • collectd::plugin::turbostat::core_c_states

  • collectd::plugin::turbostat::package_c_states

  • collectd::plugin::turbostat::system_management_interrupt

  • collectd::plugin::turbostat::digital_temperature_sensor

  • collectd::plugin::turbostat::tcc_activation_temp

  • collectd::plugin::turbostat::running_average_power_limit

  • collectd::plugin::turbostat::logical_core_names

collectd::plugin::unixsock

collectd::plugin::uptime

  • collectd::plugin::uptime::interval

collectd::plugin::users

  • collectd::plugin::users::interval

collectd::plugin::uuid

  • collectd::plugin::uuid::uuid_file

  • collectd::plugin::uuid::interval

collectd::plugin::virt

Use the virt plugin to collect CPU, disk, network load, and other metrics through the libvirt API for virtual machines on the host.

Table 23. virt parameters
Parameter Type

connection

String

refresh_interval

Hash

domain

String

block_device

String

interface_device

String

ignore_selected

Boolean

plugin_instance_format

String

hostname_format

String

interface_format

String

extra_stats

String

Example configuration:
ExtraConfig:
    collectd::plugin::virt::plugin_instance_format: name
Additional resources

For more information about configuring the virt plugin, see virt.

collectd::plugin::vmem

  • collectd::plugin::vmem::verbose

  • collectd::plugin::vmem::interval

collectd::plugin::vserver

collectd::plugin::wireless

collectd::plugin::write_graphite

  • collectd::plugin::write_graphite::carbons

  • collectd::plugin::write_graphite::carbon_defaults

  • collectd::plugin::write_graphite::globals

collectd::plugin::write_http

Use the write_http output plugin to submit values to an HTTP server by using POST requests and encoding metrics with JSON, or by using the PUTVAL command.

Table 24. write_http parameters
Parameter Type

ensure

Enum[present, absent]

nodes

Hash[String, Hash[String, Scalar]]

urls

Hash[String, Hash[String, Scalar]]

manage_package

Boolean

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
      - write_http
    ExtraConfig:
        collectd::plugin::write_http::nodes:
            collectd:
                url: “http://collectd.tld.org/collectd”
                metrics: true
                header: “X-Custom-Header: custom_value"
Additional resources
  • For more information about configuring the write_http plugin, see write_http.

collectd::plugin::write_kafka

Use the write_kafka plugin to send values to a Kafka topic. Configure the write_kafka plugin with one or more topic blocks. For each topic block, you must specify a unique name and one Kafka producer. You can use the following per-topic parameters inside the topic block:

Table 25. write_kafka parameters
Parameter Type

kafka_hosts

Array[String]

kafka_port

Integer

topics

Hash

properties

Hash

meta

Hash

Example configuration:
parameter_defaults:
    CollectdExtraPlugins:
       - write_kafka
    ExtraConfig:
      collectd::plugin::write_kafka::kafka_hosts:
        - nodeA
        - nodeB
      collectd::plugin::write_kafka::topics:
        some_events:
          format: JSON
Additional resources:

For more information about how to configure the write_kafka plugin, see write_kafka.

collectd::plugin::write_log

  • collectd::plugin::write_log::format

collectd::plugin::zfs_arc

  • None