django_http_requests_before_middlewares_total，Shouldn't this indicator be a line that keeps rising?

[luoxiaopeng]

django_http_requests_before_middlewares_total is a counter, but why sometimes it goes down？

Maybe this is not issue, just discuss.

[luoxiaopeng]

and also django_http_requests_latency_including_middlewares_seconds_sum

[luoxiaopeng]

Maybe this reason: https://github.com/korfuri/django-prometheus/blob/master/documentation/exports.md#exporting-metrics-in-a-wsgi-application-with-multiple-processes-globally

[radekska]

Hey @luoxiaopeng setting the PROMETHEUS_MULTIPROC_DIR=/tmp environment variable for my dockerized Django app solves the mentioned issue!