Hello dear Druid Users.
We have recently decided to upgrade our cluster performance monitoring from Tranquility to Kafka, but with no success. I personally feel like I am missing some basic knowledge regarding how the overall setup should look with Kafka.
Our old setup was a Tranquility server, started up with a spec (.json) for the metrics dataSource.
the Tranquility server would then send realtime tasks to our middleManagers and the Metrics dataSource would be filled with data, which we would then visualize with superset.
(this is already half magical to me...)
Now, I am trying to replace the Tranquility server with a Kafka server, but can't find what I am literally suppose to do. whats the steps?
I followed the docs at
here, got a supervisor going (presented on the overlord GUI) and lots of "index_kafka_metrics" tasks. no idea what actuly happened in the background, and the Metrics dataSource did not update.
I tried the emitter option
here, I've set it on a single broker in the cluster. all I got was this in the logs:
```
Skipping druid.emitter.kafka.metric.topic property: one of it's prefixes is also used as a property key. Prefix: druid
Skipping druid.emitter.kafka.alert.topic property: one of it's prefixes is also used as a property key. Prefix: druid
Skipping druid.emitter.kafka.bootstrap.servers property: one of it's prefixes is also used as a property key. Prefix: druid
Skipping druid.emitter.kafka.producer.config property: one of it's prefixes is also used as a property key. Prefix: druid
```
I dont know if the docs are lacking the pure basics of what needs to be done to get the Metrics dataSource up and filling, or if I'm clouded by spending to much time on this without really understanding it.
A helping hand will be super appreciated!
let me know what information should I share.
best regards,
Avinoam