Message volume decrease and CPU increase.

[PDelVal]

I am using prom-kafka-adapter:1.7.0 as a 3-replica pod in Openshfit to send metrics from Prometheus to Kafka.

When there is a fail in network between prom-kafka-adapter and Kafka, Queue from adapter fill (Error message: Queue Full Error) and CPU increased. After network connectivity restored, the CPU remain High and normal volume of messages is not recovered until the pods are restarted.

This is my configuration for remote_write on prometheus:

      remoteWrite:
      - url: http://prometheus-kafka-adapter.prom-kafka-adapter.svc.cluster.local:8080/receive
        writeRelabelConfigs:
        - action: labeldrop
          regex: service|endpoint|prometheus|prometheus_replica|cluster|id|instance`

Is someone having the same issues?

[palmerabollo]

Not sure about it. Were you able to fix the issue @PDelVal? Could you please try with the latest 1.9.0 version?