Adrian Muraru

Let me time this and get back - my observation so far is that on two different clusters with same number of consumer groups, the overall scrape time is 3x...

Renamed the issue title to better capture the feature request

@baluchicken @stoader what's your thought on adding a new field in KafkaCluster CRD to define clusterDefaultBrokerConfig ? Something similar to clusterImage but for all broker configs?

yaml anchors sounds like a neat idea. Do you know if kubectl apply understands though for example?

Fair point - we can put this behind a feature flag, right @ctrlaltluc ?

IIUC the proposal is to have the KafkaCluster CRD extended to support an optional flag for this. Something like `EnableHttp10ForHealthCheckPort` (probably shorter :) ) in: https://github.com/banzaicloud/koperator/blob/1975aebe9020f2009a96ffc416fbe7db7b2a1623/api/v1beta1/kafkacluster_types.go#L268-L297

> I wonder if this flag can be enabled or not through the [CommandLineArgs](https://github.com/banzaicloud/koperator/blob/1975aebe9020f2009a96ffc416fbe7db7b2a1623/api/v1beta1/kafkacluster_types.go#L296) Just checked that but envoy doesn't support enabling this via command line arguments

I did a quick test in my environment and switched to `adoptopenjdk/openjdk11:x86_64-alpine-jre-11.0.5_10` base image https://github.com/wurstmeister/kafka-docker/compare/master...amuraru:java11 This seems to work. Would it make sense to add support for java11 in https://github.com/wurstmeister/kafka-docker/pull/536...

This docker image is used by BanzaiCloud/kafka-operator for k8s. I've put a tutorial on how you can install locally in a kind cluster: https://github.com/amuraru/k8s-kafka-operator

thanks for reporting this @cunningr Would you try to file a PR please to fix this behaviour? thanks in advance