Ambud

Here's the bug in your setup: `bin/zookeeper-shell.sh localhost:2181` vs `-zookeeper 127.0.0.1:2181/cluster1` So either Kafka brokers should be using the same znode as kafka stats or the other way around.

Could you please confirm that the zk path used in kafkastats is the same as what the broker is using? From the output posted in the original question that is...

We haven't tested Dr.K in operations on compose so I am not sure if it will behave as expected. The two exception posted are pointing to two different issues. The...

Dr. Kafka is designed to manage 10s of clusters. Ideally deployments should separate those clusters but it's not necessary since the replication factor of the stats topic should ideally be...

@yhzhao could you please check the Kafka message format of the broker stats topic? We are expecting at least 0.10.0 for Dr.Kafka 0.2.4.3 above. Starting 0.2.4.3 we switched to using...

@BrianGallew could you please open a PR with the above diff? The issue to investigate would be why are there no stats written to certain partitions of the stats topic.

Dr. Kafka wasn't designed to have single node multi-broker operations so ideally brokers should be well distributed across nodes with 1 broker per node.

Could you please elaborate your question?

Are you using transactions in Kafka? if not consumer will be replaying messages.

Do you see the same problem when building from main branch?