search

elastic / apm-server

https://www.elastic.co/guide/en/apm/guide/current/index.html
Other
1.22k stars 523 forks source link

Feature: export prometheus metrics for monitoring/autoscaling #4299

Open dobesv opened 4 years ago

dobesv commented 4 years ago

We use prometheus metrics for auto-scaling purposes in our kubernetes cluster. This allows us to scale webservers on metrics like requests per second or queue length.

It would be nice if we could scale apm-server this way as it gives better results that scaling purely on CPU.

Currently we scale just on CPU and it seems like we are sometimes having capacity issues.

axw commented 4 years ago

@dobesv thanks for opening the issue!

This has come up (indirectly, as apm-server is based on beats) at https://github.com/elastic/beats/issues/7432. That ended up getting closed, but we'll have another chat with the beats team to see if anything has changed, and whether we want to take this on now.

In the mean time, you might want to take a look at https://github.com/trustpilot/beat-exporter. It doesn't appear to support apm-server right now, but it shouldn't be too difficult to modify it to do so.