GKE 上的 Prometheus 监控 Strimzi Kafka - 如何获取 Prometheus Pod IP

Question

I'm trying to deploy Prometheus on GKE to monitor an existing Strimzi Kafka GKE cluster, and am facing issues. (ref - https://strimzi.io/docs/operators/latest/deploying.html#proc-metrics-deploying-prometheus-operator-str )

Here is what is done:

• created a namespace - monitoring, while Kafka is deployed in namespace - kafka
• modified the kafka-deployment.yaml to include metricsConfig and KafkaExporter as specified in file https://github.com/strimzi/strimzi-kafka-operator/tree/0.26.0/examples/metrics/kafka-metrics.yaml

here are the changes:

apiVersion: kafka.strimzi.io/v1beta2
kind: Kafka
metadata:
  name: prom-graf #1
spec:
  kafka:
    version: 3.0.0
    ........
     metricsConfig:
      type: jmxPrometheusExporter
      valueFrom:
        configMapKeyRef:
          name: kafka-metrics
          key: kafka-metrics-config.yml 
     zookeeper:
      ......
      metricsConfig:
      type: jmxPrometheusExporter
      valueFrom:
        configMapKeyRef:
          name: kafka-metrics
          key: zookeeper-metrics-config.yml    
    entityOperator: #11
    topicOperator: {}
    userOperator: {}
  kafkaExporter:
    topicRegex: ".*"
    groupRegex: ".*" 
 

Command to apply changes to Kafka cluster:

kubectl apply -f kafka-deployment.yaml -n kafka

2. created the Prometheous operator

command :
curl -s https://raw.githubusercontent.com/coreos/prometheus-operator/master/bundle.yaml > bundle.yaml

changed the namespace to 'monitoring' in bundle.yaml and created the Prometheous operator

kubectl create -f bundle.yaml -n monitoring 

Ran the following commands to create the Prometheus & Grafana on GKE: files are the same as in github - https://github.com/strimzi/strimzi-kafka-operator/tree/0.26.0/examples/metrics , although i've mad the namespace changes

kubectl apply -f strimzi-pod-monitor.yaml  
(changed namespace to kafka, since kafka is deployed in namespace kafka)
kubectl apply -f prometheus-rules.yaml -n monitoring
kubectl apply -f prometheus.yaml -n monitoring
kubectl apply -f grafana.yaml - monitoring

- created secret from file - prometheus-additional.yaml
kubectl create secret generic additional-scrape-configs --from-file=prometheus-additional.yaml -n monitoring 

When i inspect the object created on namespace - monitoring, here is what is listed:

Karans-MacBook-Pro:prometheus-yamls karanalang$ kc get all,prometheus -n monitoring
NAME                                       READY   STATUS    RESTARTS   AGE
pod/grafana-5ccfb68647-55mzk               0/1     Running   0          21s
pod/prometheus-operator-85d84bb848-57lrn   1/1     Running   0          46m

NAME                          TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)    AGE
service/grafana               ClusterIP   10.100.3.10   <none>        3000/TCP   21s
service/prometheus-operator   ClusterIP   None          <none>        8080/TCP   46m

NAME                                  READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/grafana               0/1     1            0           24s
deployment.apps/prometheus-operator   1/1     1            1           46m

NAME                                             DESIRED   CURRENT   READY   AGE
replicaset.apps/grafana-5ccfb68647               1         1         0       23s
replicaset.apps/prometheus-operator-85d84bb848   1         1         1       46m

NAME                                          VERSION   REPLICAS   AGE
prometheus.monitoring.coreos.com/prometheus             3          44m

Question is - how do i get the hostname for the Prometheus install (since i need this to add as source in Grafana)? Somehow, i don't see the prometheus pod, should that not be created?

any ideas how to debug/fix this? tia !

Answer 1

Assuming that the Prometheus pods did start, their collective hostnames would be found via service discovery like prometheus.monitoring.svc.cluster.local

https://cloud.google.com/kubernetes-engine/docs/concepts/service-discovery

You might also be interesting in exposing Prometheus itself

Answer 2

this issue is resolved, file prometheus-additional.yaml is already a secret yaml file, and the secret should be created using command

kubectl create -f prometheus-additional.yaml

instead of :

kubectl create secret generic additional-scrape-configs --from-file=prometheus-additional.yaml

ref - https://issueexplorer.com/issue/strimzi/strimzi-kafka-operator/5873

seems the format of the file is changed in the recent strimzi releases