What's the best practice to use metricbeat for prometeus protocol collection on Kubernetes cluster

We are deploying metricbeat as a daemon set to collect vm、docker metrics. It works well.

If we want metricbeat to collect Pulsar、Flink(they are Deploy on k8s)'s prometheus metrics.

Currently, I found that metricbeat promtheus module only support config ip address, which can changed all the time.

I was wondering if it will be nice for metricbeat to support service discovery mechanism ?

Additional, StatefulSet on k8s can be supported by deploy a single metricbeat and config these domains on metricbeat. For statefulset, is this better or service discover better?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.