Doc needed: How to use this with apache daemon-mode processes?

[uriklagnes]

Hi,
I'm looking for documentation on how to use this with > 1 apache daemon mode processes.

Will all the metrics across all processes be aggregated by the main apache daemon and served at its port @ url /metrics?
Also, notes on the performance impact (roughly) per process.

[geobeau]

Not sure if it's what you are asking but I made a PR about updating doc for multiprocess run: aeb891b

It's for uwsgi but some of it can still be used for Apache.

[korfuri]

I completely missed that PR, sorry! I just merged it.
@uriklagnes I assume that @geobeau's doc answers your question, please reopen this issue if it doesn't.

[uriklagnes]

@geobeau , Will everything simply work seamlessly on multi-process setups?

i.e., if we have > 1 worker processes, will the counters returned by django-prometheus be aggregate across all workers?

I found some documentation: https://github.com/korfuri/django-prometheus/blob/master/documentation/exports.md

It's quite well documented for situations with multiple processes (that change a lot) under uWSGI, but not under Apache or Nginx. Is there any way we can add thos?

[geobeau]

Well I only used multiprocesses for uWSGI so I'm not sure about Apache/Nginx. If you set the env prometheus_multiproc_dir=/path/to/django_metrics in your django process it should work out of the box. So I would recommend that first you try just with that, then address the scalability issues if you have some.