jmx-monitoring-stacks
jmx-monitoring-stacks copied to clipboard
zookeeper avg request latency is multiplied by ticktime
In the grafana dashboard for Zookeeper, I see that the latency metrics are being multiplied by zookeeper ticktime. This has the effect of multiplying ZK request latency by 3000 (the default ticktime in my ZK version).
That change was introduced as part of the following PR: https://github.com/confluentinc/jmx-monitoring-stacks/pull/22
What is the rationale for doing this?
Zookeeper latency is (was?) measured in ticks, that's why this avg latency was 0 most of the time. If it has changes we can update it for later versions of zookeeper (or wait for KRaft :))
fixed with: https://github.com/confluentinc/jmx-monitoring-stacks/pull/172