Missing metrics

[Azbesciak]

Hello, I think it is not the issue of this configuration, but many people use this config and might face the same issue.
I wanted to monitor the average request response time, quintiles, errors count, request count, most often IPs, etc.
I found that it is possible with for example http_request_total - but it is not available in the Prometheus.
I read somewhere that there is promhttp_metric_handler_requests_total, and indeed - there is something like this, but returns something different.

How to solve this problem? I see that http_request_total is not something unique and commonly used, for example, https://blog.softwaremill.com/practical-monitoring-with-prometheus-ee09a1dd5527.

Thanks!

[nightah]

@Azbesciak can you clarify you want to see those metrics from what specific service, Prometheus itself or?

[Azbesciak]

@nightah from all containers

[nightah]

So the only thing that exists within the stack is prometheus_http_requests_total which returns the values against a specific URI/path and the associated http status code.

I'd suggest you look into a prometheus exporter specifically for your reverse proxy (nginx, Traefik, HAProxy) and utilise that to get the information you seek.

[nightah]

@Azbesciak I'm going to close off this issue, if you have any other questions please don't hesitate to open another issue.