linkedin
Burrow doesn't work at all even if just 1 Kafka cluster out of several other clusters are not reachable.
Burrow is starting and seem to be working, but in reality it is not providing any consumer metrics of any Kafka cluster. It's because I have about 10 Kafka clusters and 2-3 of them are not working (e.g. for testing purposes).
If I remove those testing clusters from the burrow and all other remaining clusters are working fine - Burrow is also working fine.
Am I missing something or is it intended behavior?
Still revelant after more than a year...... :(
Imagine this:
test Kafka "cluster" of 1 server
preproduction: Kafka cluster of 3 servers
production: Kafka cluster of 3 servers
Sometimes you want to upgrade test Kafka cluster, reboot it or power ir down when not in use. However, Burrow stops working COMPLETELLY if test cluster is down. WTF is this behaviour, who came up with this design in the first place??
{"level":"error","ts":1647443816.1310833,"msg":"failed to start client","type":"module","coordinator":"cluster","class":"kafka","name":"kafka_test","error":"kafka: client has run out of available brokers to talk to (Is your cluster reachable?)"}
How am I suppossed to use Burrow in HA/Centralised fashion? Deploy a separate Burrow instance per cluster??