Monitoring support

[zamazan4ik]

Is your feature request related to a problem? Please describe. I want to monitor Quickwit's metrics and health checks. And I want to do it with a standard for the industry tooling like Prometheus, Grafana, etc.

Describe the solution you'd like

• Integrate Prometheus metrics into Quickwit
• Provide ready to use Grafana dashboards
• Add notes about alerting setup (possibly via alert-manager or something similar)
• Update documentation how to use it

Describe alternatives you've considered Just do not introduce monitoring

[fulmicoton]

Hello @zamazan4ik ,

We haven't document it much, but we do a set of prometheus metrics. You can access them on /metrics

[fmassot]

We have just added the metrics in the docs: https://github.com/quickwit-oss/quickwit/blob/main/docs/reference/metrics.md

[zamazan4ik]

@fmassot Thanks! For the users would be much easier to read it the information (at least some general info) directly from the documentation on the site. And there you can add a link to the metrics.md page.

[guilload]

Unfortunately, publishing the docs from our main repo to our website repo is not automated (yet). The metrics page will eventually be available on quickwit.io/docs, probably when we release 0.4 in a few weeks.

[fmassot]

Unfortunately, publishing the docs from our main repo to our website repo is not automated (yet)

It is... but it was broken as one page had a broken link. It's fixed now: https://quickwit.io/docs/main-branch/reference/metrics

[fmassot]

@zamazan4ik I opened a dedicated issue for the indexer monitoring #2160

[ppaulweber]

@fmassot @fulmicoton is there an official Prometheus write API endpoint in quickwit?

Since the stream of logs is covered and the persisting of trace information throught he dedicated gRPC OTEL endpoint I was wandering if quickwit already (or will) support persisting of (e.g. Prometheus) metric information as well through a dedicated API ingest endpoint?

[fmassot]

@ppaulweber support for metrics is planned for 2024. We planned to work on two topics: storage optimized for time series + write/read API compatible with standards (like Prometheus). We don't have yet an accurate roadmap on this.

[ppaulweber]

@fmassot thanks for this quick update and info.