deploy

Deploy Prometurbo

It is recommended to deploy Prometurbo via operator. The following are the steps to install Prometurbo Operator through OpenShift OperatorHub, create a Prometurbo instance, and then create PrometheusQueryMapping and PrometheusServerConfig custom resources to be consumed by Prometurbo.

Install Prometurbo Operator through OperatorHub in OpenShift

• Create a project (namespace) for your Prometurbo deployment. For example, the following YAML file creates a turbo namespace:

  apiVersion: v1
  kind: Namespace
  metadata:
    name: turbo 

• Navigate to the OperatorHub section in the OpenShift console, select the project created above, search for Prometurbo in the search bar, and select the Certified Prometurbo Operator:

  

• Make sure the stable channel is selected, and install the operator:

  

  

• After the install, the following resources are created:

  Name	Kind	Status	API version
  prometurbo-operator.vx.x.x	ClusterServiceVersion	Created	operators.coreos.com/v1alpha1
  prometurbos.charts.helm.k8s.io	CustomResourceDefinition	Present	apiextensions.k8s.io/v1
  prometurbo-operator	ServiceAccount	Present	core/v1
  prometurbo-operator.vx.x.x-xxxxxxxx	ClusterRole	Created	rbac.authorization.k8s.io/v1
  prometurbo-operator.vx.x.x-xxxxxxxx	ClusterRoleBinding	Created	rbac.authorization.k8s.io/v1

  Note that ClusterRole is created for prometurbo-operator such that it can have the permission to create the necessary ClusterRole for Prometurbo instances.

Create a Prometurbo instance

• Once the operator is installed, navigate to the Installed Operators section in the OpenShift console. Select the Prometurbo operator and click Create Instance. Follow the prompts to configure the instance.

• The following table lists the configuration parameters for prometurbo:

  Parameter	Description	Default Value
  image.prometurboRepository	The prometurbo container image repository.	icr.io/cpopen/turbonomic/prometurbo
  image.prometurboTag	The prometurbo container image tag.	release version, such as 8.8.4
  image.turbodifRepository	The turbodif container image repository.	icr.io/cpopen/turbonomic/turbodif
  image.turbodifTag	The turbodif container image tag.	release version, such as 8.8.4
  image.pullPolicy	Specify either IfNotPresent, or Always.	IfNotPresent
  targetName	A unique name to identify this target.	Prometheus
  serviceAccountName	The name of the serviceAccount used by prometurbo pod.	prometurbo
  roleName	The name of the clusterrole bound to the above service account.	prometurbo
  roleBinding	The name of the clusterrolebinding that binds the above service account to the above cluster role.	prometurbo-binding
  restAPIConfig.turbonomicCredentialsSecretName	The name of the secret that contains Turbonomic server OAuth2 clientid and clientsecret. Required if not using cleartext username and password, and not taking the default secret name. If the secrect does not exist, prometurbo falls back to cleartext username and password.	turbonomic-credentials
  restAPIConfig.opsManagerUserName	The username to login to the Turbonomic server. Required if not using secret.
  restAPIConfig.opsManagerPassword	The password to login to the Turbonomic server. Required if not using secret.
  serverMeta.version	Turbonomic server version.	release version, such as 8.8.4
  serverMeta.turboServer	Turbonomic server URL.
  args.logginglevel	Logging level of prometurbo.	2
  args.ignoreCommodityIfPresent	Specify whether to ignore merging commodity when a commodity of the same type already exists in the server.	false

• The following is a sample Prometurbo resource YAML file:

  apiVersion: charts.helm.k8s.io/v1
  kind: Prometurbo
  metadata:
    name: prometurbo-release
    namespace: turbo
  spec:
    image:
      prometurboTag: 8.8.4
      turbodifTag: 8.8.4
      pullPolicy: Always
    restAPIConfig:
      opsManagerPassword: administrator
      opsManagerUserName: administrator
    serviceAccountName: prometurbo-xl-ember
    roleBinding: prometurbo-binding-xl-ember
    roleName: prometurbo-xl-ember
    serverMeta:
      turboServer: 'https://9.46.114.249'
      version: 8.8.4
    targetName: OCP411-FYRE-IBM

• Verify the install by making sure that prometurbo pods are up and running:

$ oc -n turbo get po | grep prometurbo
prometurbo-operator-6ffc566f4c-lwbcn       1/1     Running   0          5d23h
prometurbo-release-744947bb94-kqc2b        2/2     Running   0          5d20h

• Verify that proper ClusterRole is created for Prometurbo. For example, given the above YAML, a prometurbo-xl-ember ClusterRole should be created, with proper rules to access the prometheusquerymappings and prometheusserverconfigs resources:

  

Enable turbo-metrics

• Install Custom Resource Definitions

$ oc create -f https://raw.githubusercontent.com/turbonomic/turbo-metrics/main/config/crd/bases/metrics.turbonomic.io_prometheusquerymappings.yaml
$ oc create -f https://raw.githubusercontent.com/turbonomic/turbo-metrics/main/config/crd/bases/metrics.turbonomic.io_prometheusserverconfigs.yaml

• Create PrometheusQueryMapping and PromethesServerConfig resources. These custom resources allow you to define configurations for Prometurbo to consume:

  • Create a PrometheusQueryMapping resource to specify how Prometurbo should map Prometheus queries to Turbonomic entities. The following is a sample of PrometheusQueryMapping for metrics exposed by istio exporter:

    apiVersion: metrics.turbonomic.io/v1alpha1
    kind: PrometheusQueryMapping
    metadata:
      labels:
        mapping: istio
      name: istio
    spec:
      entities:
        - type: application
          metrics:
            - type: responseTime
              queries:
                - type: used
                  promql: 'rate(istio_request_duration_milliseconds_sum{request_protocol="http",response_code="200",reporter="destination"}[1m])/rate(istio_request_duration_milliseconds_count{}[1m]) >= 0'
            - type: transaction
              queries:
                - type: used
                  promql: 'rate(istio_requests_total{request_protocol="http",response_code="200",reporter="destination"}[1m]) > 0'
            - type: responseTime
              queries:
                - type: used
                  promql: 'rate(istio_request_duration_milliseconds_sum{request_protocol="grpc",grpc_response_status="0",response_code="200",reporter="destination"}[1m])/rate(istio_request_duration_milliseconds_count{}[1m]) >= 0'
            - type: transaction
              queries:
                - type: used
                  promql: 'rate(istio_requests_total{request_protocol="grpc",grpc_response_status="0",response_code="200",reporter="destination"}[1m]) > 0'
          attributes:
            - name: ip
              label: instance
              matches: \d{1,3}(?:\.\d{1,3}){3}(?::\d{1,5})??
              isIdentifier: true
            - name: namespace
              label: destination_service_namespace
            - name: service
              label: destination_service_name

  • Create a PrometheusServerConfig resource to specify configuration options for the Prometheus server. The following is an example of a PrometheusServerConfig resource which specifies the location of the Prometheus server, and a label selector to exclude jmx-tomcat PrometheusQueryMapping resource:

    apiVersion: metrics.turbonomic.io/v1alpha1
    kind: PrometheusServerConfig
    metadata:
      name: prometheusserverconfig-emptycluster
    spec:
      address: http://prometheus.istio-system:9090
      bearerToken:
          secretKeyRef:
            key: authorizationToken
            name: ocp-thanos-authorization-token
      clusters:
        - queryMappingSelector:
            matchExpressions:
              - key: mapping
                operator: NotIn
                values:
                  - jmx-tomcat

• Prometurbo should now be ready to consume metrics from Prometheus server. Check prometurbo logs to verify. For example:

  I0328 18:42:04.003329 1 provider.go:60] Discovered 4 PrometheusQueryMapping resources.
  I0328 18:42:04.007689 1 provider.go:71] Discovered 2 PrometheusServerConfig resources.
  I0328 18:42:04.007903 1 serverconfig.go:19] Loading PrometheusServerConfig turbo-community/prometheusserverconfig-emptycluster.
  I0328 18:42:04.007927 1 client.go:68] Creating client for Prometheus server: http://prometheus.istio-system:9090
  I0328 18:42:04.007935 1 serverconfig.go:36] There are 1 PrometheusQueryMapping resources in namespace turbo-community
  I0328 18:42:04.007943 1 serverconfig.go:19] Loading PrometheusServerConfig turbo/prometheusserverconfig-singlecluster.
  I0328 18:42:04.007947 1 client.go:68] Creating client for Prometheus server: http://prometheus.istio-system:9090
  I0328 18:42:04.007950 1 serverconfig.go:36] There are 3 PrometheusQueryMapping resources in namespace turbo
  I0328 18:42:04.008048 1 clusterconfig.go:39] Excluding turbo/jmx-tomcat.