Kafka-UI monitoring

[timchenko-a]

Hi, what is a proper way to monitor Kafka-UI? I'm deploying it to k8s with the provided Helm chart, but seems like there is no pod/service monitor available and I haven't found anything related to metrics of Kafka UI itself in the documentation. Does it expose any metrics we can collect?

[Haarolean]

Hi, to better understand your inquiry: what kind of ui-related metrics do you need?

Right now you can monitor some stuff via these endpoints:
/actuator/health
/actuator/info

[timchenko-a]

Will try to be specific :)
So I use audit logging feature for one of my clusters which I access via Kafka-UI. And I had some kind of an issue when audit logs were generated by Kafka-UI (I can see them in the logs), but were not send to the Kafka topic. Restart of Kafka-UI kind of solved the issue though.
So I'm curious if it's possible to spot such issues somehow using metrics. E.g. Kafka-UI will expose a metric with number of generated audit events and a Kafka Producer metric, reporting number of messages send to Kafka. Comparing these metrics will allow to create an alert if some audit events were not send to Kafka for whatever reason. Does it make any sense?

[Haarolean]

Guess we could expose an audit 'readiness' into /actuator/health. Exposing metrics on this looks like an overkill, at least at this moment.

[timchenko-a]

Got it, thanks. I thought maybe backend part exposes some generic metrics, like requests/errors/latency, and in this case it's just a matter of adding a couple of additional metrics.