New to KubeDB? Please start here.

Monitoring Kafka Using Prometheus operator

Prometheus operator provides simple and Kubernetes native way to deploy and configure Prometheus server. This tutorial will show you how to use Prometheus operator to monitor Kafka database deployed with KubeDB.

Before You Begin

• At first, you need to have a Kubernetes cluster, and the kubectl command-line tool must be configured to communicate with your cluster. If you do not already have a cluster, you can create one locally by using kind.

• To learn how Prometheus monitoring works with KubeDB in general, please visit here.

• We need a Prometheus operator instance running. If you don’t already have a running instance, you can deploy one using this helm chart here.

• To keep Prometheus resources isolated, we are going to use a separate namespace called monitoring to deploy the prometheus operator helm chart. Alternatively, you can use --create-namespace flag while deploying prometheus. We are going to deploy database in demo namespace.

  $ kubectl create ns monitoring
  namespace/monitoring created
  
  $ kubectl create ns demo
  namespace/demo created
  

Note: YAML files used in this tutorial are stored in docs/examples/kafka folder in GitHub repository kubedb/docs.

Find out required labels for ServiceMonitor

We need to know the labels used to select ServiceMonitor by a Prometheus crd. We are going to provide these labels in spec.monitor.prometheus.serviceMonitor.labels field of Kafka crd so that KubeDB creates ServiceMonitor object accordingly.

At first, let’s find out the available Prometheus server in our cluster.

$ kubectl get prometheus --all-namespaces
NAMESPACE    NAME                                    VERSION   DESIRED   READY   RECONCILED   AVAILABLE   AGE
monitoring   prometheus-kube-prometheus-prometheus   v2.42.0   1         1       True         True        2d23h

If you don’t have any Prometheus server running in your cluster, deploy one following the guide specified in Before You Begin section.

Now, let’s view the YAML of the available Prometheus server prometheus in monitoring namespace.

$ kubectl get prometheus -n monitoring prometheus-kube-prometheus-prometheus -o yaml
apiVersion: monitoring.coreos.com/v1
kind: Prometheus
metadata:
  annotations:
    meta.helm.sh/release-name: prometheus
    meta.helm.sh/release-namespace: monitoring
  creationTimestamp: "2023-03-27T07:56:04Z"
  generation: 1
  labels:
    app: kube-prometheus-stack-prometheus
    app.kubernetes.io/instance: prometheus
    app.kubernetes.io/managed-by: Helm
    app.kubernetes.io/part-of: kube-prometheus-stack
    app.kubernetes.io/version: 45.7.1
    chart: kube-prometheus-stack-45.7.1
    heritage: Helm
    release: prometheus
  name: prometheus-kube-prometheus-prometheus
  namespace: monitoring
  resourceVersion: "638797"
  uid: 0d1e7b8a-44ae-4794-ab45-95a5d7ae7f91
spec:
  alerting:
    alertmanagers:
    - apiVersion: v2
      name: prometheus-kube-prometheus-alertmanager
      namespace: monitoring
      pathPrefix: /
      port: http-web
  enableAdminAPI: false
  evaluationInterval: 30s
  externalUrl: http://prometheus-kube-prometheus-prometheus.monitoring:9090
  hostNetwork: false
  image: quay.io/prometheus/prometheus:v2.42.0
  listenLocal: false
  logFormat: logfmt
  logLevel: info
  paused: false
  podMonitorNamespaceSelector: {}
  podMonitorSelector:
    matchLabels:
      release: prometheus
  portName: http-web
  probeNamespaceSelector: {}
  probeSelector:
    matchLabels:
      release: prometheus
  replicas: 1
  retention: 10d
  routePrefix: /
  ruleNamespaceSelector: {}
  ruleSelector:
    matchLabels:
      release: prometheus
  scrapeInterval: 30s
  securityContext:
    fsGroup: 2000
    runAsGroup: 2000
    runAsNonRoot: true
    runAsUser: 1000
  serviceAccountName: prometheus-kube-prometheus-prometheus
  serviceMonitorNamespaceSelector: {}
  serviceMonitorSelector:
    matchLabels:
      release: prometheus
  shards: 1
  version: v2.42.0
  walCompression: true
status:
  availableReplicas: 1
  conditions:
  - lastTransitionTime: "2023-03-27T07:56:23Z"
    observedGeneration: 1
    status: "True"
    type: Available
  - lastTransitionTime: "2023-03-30T03:39:18Z"
    observedGeneration: 1
    status: "True"
    type: Reconciled
  paused: false
  replicas: 1
  shardStatuses:
  - availableReplicas: 1
    replicas: 1
    shardID: "0"
    unavailableReplicas: 0
    updatedReplicas: 1
  unavailableReplicas: 0
  updatedReplicas: 1

Notice the spec.serviceMonitorSelector section. Here, release: prometheus label is used to select ServiceMonitor crd. So, we are going to use this label in spec.monitor.prometheus.serviceMonitor.labels field of Kafka crd.

Deploy Kafka with Monitoring Enabled

At first, let’s deploy a Kafka database with monitoring enabled. Below is the Kafka object that we are going to create.

apiVersion: kubedb.com/v1alpha2
kind: Kafka
metadata:
  name: kafka
  namespace: demo
spec:
  enableSSL: true
  tls:
    issuerRef:
      apiGroup: cert-manager.io
      name: kafka-ca-issuer
      kind: Issuer
  replicas: 3
  version: 3.4.0
  storage:
    accessModes:
      - ReadWriteOnce
    resources:
      requests:
        storage: 1Gi
    storageClassName: standard
  monitor:
    agent: prometheus.io/operator
    prometheus:
      exporter:
        port: 9091
      serviceMonitor:
        labels:
          release: prometheus
        interval: 10s
  storageType: Durable
  terminationPolicy: WipeOut

Here,

• monitor.agent: prometheus.io/operator indicates that we are going to monitor this server using Prometheus operator.
• monitor.prometheus.serviceMonitor.labels specifies that KubeDB should create ServiceMonitor with these labels.
• monitor.prometheus.interval indicates that the Prometheus server should scrape metrics from this database with 10 seconds interval.

Let’s create the kafka object that we have shown above,

$ kubectl create -f https://github.com/kubedb/docs/raw/v2024.1.31/docs/examples/kafka/monitoring/kf-with-monirtoring.yaml
kafkas.kubedb.com/kafka created

Now, wait for the database to go into Running state.

$ kubectl get kf -n demo kafka
NAME    TYPE                  VERSION   STATUS   AGE
kafka   kubedb.com/v1alpha2   3.4.0     Ready    2m24s

KubeDB will create a separate stats service with name {Kafka crd name}-stats for monitoring purpose.

$ kubectl get svc -n demo --selector="app.kubernetes.io/instance=kafka"
NAME          TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)                       AGE
kafka-pods    ClusterIP   None            <none>        9092/TCP,9093/TCP,29092/TCP   3m22s
kafka-stats   ClusterIP   10.96.235.251   <none>        9091/TCP                      3m19s

Here, kafka-stats service has been created for monitoring purpose.

Let’s describe this stats service.

$ kubectl describe svc -n demo kafka-stats
Name:              kafka-stats
Namespace:         demo
Labels:            app.kubernetes.io/component=database
  app.kubernetes.io/instance=kafka
  app.kubernetes.io/managed-by=kubedb.com
  app.kubernetes.io/name=kafkas.kubedb.com
  kubedb.com/role=stats
Annotations:       monitoring.appscode.com/agent: prometheus.io/operator
Selector:          app.kubernetes.io/instance=kafka,app.kubernetes.io/managed-by=kubedb.com,app.kubernetes.io/name=kafkas.kubedb.com
Type:              ClusterIP
IP Family Policy:  SingleStack
IP Families:       IPv4
IP:                10.96.235.251
IPs:               10.96.235.251
Port:              metrics  9091/TCP
TargetPort:        metrics/TCP
Endpoints:         10.244.0.117:56790,10.244.0.119:56790,10.244.0.121:56790
Session Affinity:  None
Events:            <none>

Notice the Labels and Port fields. ServiceMonitor will use this information to target its endpoints.

KubeDB will also create a ServiceMonitor crd in demo namespace that select the endpoints of kafka-stats service. Verify that the ServiceMonitor crd has been created.

$ kubectl get servicemonitor -n demo
NAME          AGE
kafka-stats   4m49s

Let’s verify that the ServiceMonitor has the label that we had specified in spec.monitor section of Kafka crd.

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  creationTimestamp: "2023-03-30T07:59:49Z"
  generation: 1
  labels:
    app.kubernetes.io/component: database
    app.kubernetes.io/instance: kafka
    app.kubernetes.io/managed-by: kubedb.com
    app.kubernetes.io/name: kafkas.kubedb.com
    release: prometheus
  name: kafka-stats
  namespace: demo
  ownerReferences:
  - apiVersion: v1
    blockOwnerDeletion: true
    controller: true
    kind: Service
    name: kafka-stats
    uid: 4a95fc65-fe2c-4d9c-afdd-aa748642d6bc
  resourceVersion: "668351"
  uid: de76712d-4f51-4bab-a625-73966f4bd9f7
spec:
  endpoints:
  - bearerTokenSecret:
      key: ""
    honorLabels: true
    interval: 10s
    path: /metrics
    port: metrics
  namespaceSelector:
    matchNames:
    - demo
  selector:
    matchLabels:
      app.kubernetes.io/component: database
      app.kubernetes.io/instance: kafka
      app.kubernetes.io/managed-by: kubedb.com
      app.kubernetes.io/name: kafkas.kubedb.com
      kubedb.com/role: stats

Notice that the ServiceMonitor has label release: prometheus that we had specified in Kafka crd.

Also notice that the ServiceMonitor has selector which match the labels we have seen in the kafka-stats service. It also, target the metrics port that we have seen in the stats service.

Verify Monitoring Metrics

At first, let’s find out the respective Prometheus pod for prometheus Prometheus server.

$ kubectl get pod -n monitoring -l=app.kubernetes.io/name=prometheus
NAME                                                 READY   STATUS    RESTARTS        AGE
prometheus-prometheus-kube-prometheus-prometheus-0   2/2     Running   8 (4h27m ago)   3d

Prometheus server is listening to port 9090 of prometheus-prometheus-kube-prometheus-prometheus-0 pod. We are going to use port forwarding to access Prometheus dashboard.

Run following command on a separate terminal to forward the port 9090 of prometheus-kube-prometheus-prometheus service which is pointing to the prometheus pod,

$ kubectl port-forward -n monitoring svc/prometheus-kube-prometheus-prometheus 9090
Forwarding from 127.0.0.1:9090 -> 9090
Forwarding from [::1]:9090 -> 9090

Now, we can access the dashboard at localhost:9090. Open http://localhost:9090 in your browser. You should see metrics endpoint of kafka-stats service as one of the targets.

Check the endpoint and service labels. It verifies that the target is our expected database. Now, you can view the collected metrics and create a graph from homepage of this Prometheus dashboard. You can also use this Prometheus server as data source for Grafana and create a beautiful dashboard with collected metrics.

Cleaning up

To clean up the Kubernetes resources created by this tutorial, run following commands

kubectl delete -n demo kf/kafka
kubectl delete ns demo

Next Steps

• Learn how to use KubeDB to run a Apache Kafka cluster here.
• Deploy dedicated topology cluster for Apache Kafka
• Deploy combined cluster for Apache Kafka
• Detail concepts of KafkaVersion object.
• Learn to use KubeDB managed Kafka objects using CLIs.
• Want to hack on KubeDB? Check our contribution guidelines.