Configuring built-in monitoring with Prometheus

This guide describes the built-in monitoring support provided by the Operator SDK using the Prometheus Operator and details usage for authors of Go-based and Ansible-based Operators.

Prometheus Operator support

Prometheus is an open-source systems monitoring and alerting toolkit. The Prometheus Operator creates, configures, and manages Prometheus clusters running on Kubernetes-based clusters, such as OKD.

Helper functions exist in the Operator SDK by default to automatically set up metrics in any generated Go-based Operator for use on clusters where the Prometheus Operator is deployed.

Exposing custom metrics for Go-based Operators

As an Operator author, you can publish custom metrics by using the global Prometheus registry from the controller-runtime/pkg/metrics library.

Prerequisites

  • Go-based Operator generated using the Operator SDK

  • Prometheus Operator, which is deployed by default on OKD clusters

Procedure

  1. In your Operator SDK project, uncomment the following line in the config/default/kustomization.yaml file:

    1. ../prometheus

  2. Create a custom controller class to publish additional metrics from the Operator. The following example declares the widgets and widgetFailures collectors as global variables, and then registers them with the init() function in the controller’s package:

    controllers/memcached_controller_test_metrics.go file

    1. package controllers
    3. import (
    4.     "github.com/prometheus/client_golang/prometheus"
    5.     "sigs.k8s.io/controller-runtime/pkg/metrics"
    6. )
    8. var (
    9.     widgets = prometheus.NewCounter(
    10.         prometheus.CounterOpts{
    11.             Name: "widgets_total",
    12.             Help: "Number of widgets processed",
    13.         },
    14.     )
    15.     widgetFailures = prometheus.NewCounter(
    16.         prometheus.CounterOpts{
    17.             Name: "widget_failures_total",
    18.             Help: "Number of failed widgets",
    19.         },
    20.     )
    21. )
    23. func init() {
    24.     // Register custom metrics with the global prometheus registry
    25.     metrics.Registry.MustRegister(widgets, widgetFailures)
    26. }

  3. Record to these collectors from any part of the reconcile loop in the main controller class, which determines the business logic for the metric:

    controllers/memcached_controller.go file

    1. func (r *MemcachedReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
    2.     ...
    3.     ...
    4.     // Add metrics
    5.     widgets.Inc()
    6.     widgetFailures.Inc()
    8.     return ctrl.Result{}, nil
    9. }

  4. Build and push the Operator:

    1. $ make docker-build docker-push IMG=<registry>/<user>/<image_name>:<tag>

  5. Deploy the Operator:

    1. $ make deploy IMG=<registry>/<user>/<image_name>:<tag>

  6. Create role and role binding definitions to allow the service monitor of the Operator to be scraped by the Prometheus instance of the OKD cluster.

    Roles must be assigned so that service accounts have the permissions to scrape the metrics of the namespace:

    config/prometheus/role.yaml role

    1. apiVersion: rbac.authorization.k8s.io/v1
    2. kind: ClusterRole
    3. metadata:
    4.   name: prometheus-k8s-role
    5.   namespace: <operator_namespace>
    6. rules:
    7.   - apiGroups:
    8.       - ""
    9.     resources:
    10.       - endpoints
    11.       - pods
    12.       - services
    13.       - nodes
    14.       - secrets
    15.     verbs:
    16.       - get
    17.       - list
    18.       - watch

    config/prometheus/rolebinding.yaml role binding

    1. apiVersion: rbac.authorization.k8s.io/v1
    2. kind: ClusterRoleBinding
    3. metadata:
    4.   name: prometheus-k8s-rolebinding
    5.   namespace: memcached-operator-system
    6. roleRef:
    7.   apiGroup: rbac.authorization.k8s.io
    8.   kind: ClusterRole
    9.   name: prometheus-k8s-role
    10. subjects:
    11.   - kind: ServiceAccount
    12.     name: prometheus-k8s
    13.     namespace: openshift-monitoring

  7. Apply the roles and role bindings for the deployed Operator:

    1. $ oc apply -f config/prometheus/role.yaml
    1. $ oc apply -f config/prometheus/rolebinding.yaml

  8. Set the labels for the namespace that you want to scrape, which enables OpenShift cluster monitoring for that namespace:

    1. $ oc label namespace <operator_namespace> openshift.io/cluster-monitoring="true"

Verification

  • Query and view the metrics in the OKD web console. You can use the names that were set in the custom controller class, for example widgets_total and widget_failures_total.

Exposing custom metrics for Ansible-based Operators

As an Operator author creating Ansible-based Operators, you can use the Operator SDK’s osdk_metrics module to expose custom Operator and Operand metrics, emit events, and support logging.

Prerequisites

  • Ansible-based Operator generated using the Operator SDK

  • Prometheus Operator, which is deployed by default on OKD clusters

Procedure

  1. Generate an Ansible-based Operator. This example uses a testmetrics.com domain:

    1. $ operator-sdk init \
    2.     --plugins=ansible \
    3.     --domain=testmetrics.com

  2. Create a metrics API. This example uses a kind named Testmetrics:

    1. $ operator-sdk create api \
    2.     --group metrics \
    3.     --version v1 \
    4.     --kind Testmetrics \
    5.     --generate-role

  3. Edit the roles/testmetrics/tasks/main.yml file and use the osdk_metrics module to create custom metrics for your Operator project:

    Example roles/testmetrics/tasks/main.yml file

    1. ---
    2. # tasks file for Memcached
    3. - name: start k8sstatus
    4.   k8s:
    5.     definition:
    6.       kind: Deployment
    7.       apiVersion: apps/v1
    8.       metadata:
    9.         name: '{{ ansible_operator_meta.name }}-memcached'
    10.         namespace: '{{ ansible_operator_meta.namespace }}'
    11.       spec:
    12.         replicas: "{{size}}"
    13.         selector:
    14.           matchLabels:
    15.             app: memcached
    16.         template:
    17.           metadata:
    18.             labels:
    19.               app: memcached
    20.           spec:
    21.             containers:
    22.             - name: memcached
    23.               command:
    24.               - memcached
    25.               - -m=64
    26.               - -o
    27.               - modern
    28.               - -v
    29.               image: "docker.io/memcached:1.4.36-alpine"
    30.               ports:
    31.                 - containerPort: 11211
    33. - osdk_metric:
    34.     name: my_thing_counter
    35.     description: This metric counts things
    36.     counter: {}
    38. - osdk_metric:
    39.     name: my_counter_metric
    40.     description: Add 3.14 to the counter
    41.     counter:
    42.       increment: yes
    44. - osdk_metric:
    45.     name: my_gauge_metric
    46.     description: Create my gauge and set it to 2.
    47.     gauge:
    48.       set: 2
    50. - osdk_metric:
    51.     name: my_histogram_metric
    52.     description: Observe my histogram
    53.     histogram:
    54.       observe: 2
    56. - osdk_metric:
    57.     name: my_summary_metric
    58.     description: Observe my summary
    59.     summary:
    60.       observe: 2

Verification

  1. Run your Operator on a cluster. For example, to use the “run as a deployment” method:

    1. Build the Operator image and push it to a registry:

      1. $ make docker-build docker-push IMG=<registry>/<user>/<image_name>:<tag>

    2. Install the Operator on a cluster:

      1. $ make install

    3. Deploy the Operator:

      1. $ make deploy IMG=<registry>/<user>/<image_name>:<tag>

  2. Create a Testmetrics custom resource (CR):

    1. Define the CR spec:

      Example config/samples/metrics_v1_testmetrics.yaml file

      1. apiVersion: metrics.testmetrics.com/v1
      2. kind: Testmetrics
      3. metadata:
      4.   name: testmetrics-sample
      5. spec:
      6.   size: 1

    2. Create the object:

      1. $ oc create -f config/samples/metrics_v1_testmetrics.yaml

  3. Get the pod details:

    1. $ oc get pods

    Example output

    1. NAME                                    READY   STATUS    RESTARTS   AGE
    2. ansiblemetrics-controller-manager-<id>  2/2     Running   0          149m
    3. testmetrics-sample-memcached-<id>       1/1     Running   0          147m

  4. Get the endpoint details:

    1. $ oc get ep

    Example output

    1. NAME                                                ENDPOINTS          AGE
    2. ansiblemetrics-controller-manager-metrics-service   10.129.2.70:8443   150m

  5. Request a custom metrics token:

    1. $ token=`oc create token prometheus-k8s -n openshift-monitoring`

  6. Check the metrics values:

    1. Check the my_counter_metric value:

      1. $ oc exec ansiblemetrics-controller-manager-<id> -- curl -k -H "Authoriza
      2. tion: Bearer $token" 'https://10.129.2.70:8443/metrics' | grep  my_counter

      Example output

      1. HELP my_counter_metric Add 3.14 to the counter
      2. TYPE my_counter_metric counter
      3. my_counter_metric 2

    2. Check the my_gauge_metric value:

      1. $ oc exec ansiblemetrics-controller-manager-<id> -- curl -k -H "Authoriza
      2. tion: Bearer $token" 'https://10.129.2.70:8443/metrics' | grep  gauge

      Example output

      1. HELP my_gauge_metric Create my gauge and set it to 2.

    3. Check the my_histogram_metric and my_summary_metric values:

      1. $ oc exec ansiblemetrics-controller-manager-<id> -- curl -k -H "Authoriza
      2. tion: Bearer $token" 'https://10.129.2.70:8443/metrics' | grep  Observe

      Example output

      1. HELP my_histogram_metric Observe my histogram
      2. HELP my_summary_metric Observe my summary