Add a prefix to Prometheus metrics so they are all in one place when looking at the full list of available metrics on the Prometheus server

[pkerney]

I am just starting to explore Prometheus/Grafana metrics with MediaMTX. I note that there is no global prefix to the returned metrics. This would allow grouping of the metrics when sorting them on the Prometheus server.
This is not a big problem, but just wondering what people think about having something like "mediamtx_" as a global prefix to the metrics.
Alternatively, it could be a configuration parameter and default to "" (nothing) so as to not break existing deployments.

[aler9]

Hello Peter, Prometheus already adds a tag that contains the source name and allows to group metrics. If you have this Prometheus configuration:

scrape_configs:
- job_name: mediamtx
  scrape_interval: 5s
  static_configs:
  - targets: ['127.0.0.1:9998']

All gathered metrics will have the job="mediamtx" tag:

hls_muxers{instance="127.0.0.1:9998", job="mediamtx"}
	0
hls_muxers_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}
	0
paths{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtmp_conns{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtmp_conns_bytes_received{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtmp_conns_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_conns{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_conns_bytes_received{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_conns_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_sessions{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_sessions_bytes_received{instance="127.0.0.1:9998", job="mediamtx"}
	0
rtsp_sessions_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}
	0
scrape_duration_seconds{instance="127.0.0.1:9998", job="mediamtx"}
	0.001741061
scrape_samples_post_metric_relabeling{instance="127.0.0.1:9998", job="mediamtx"}
	18
scrape_samples_scraped{instance="127.0.0.1:9998", job="mediamtx"}
	18
scrape_series_added{instance="127.0.0.1:9998", job="mediamtx"}
	0
srt_conns{instance="127.0.0.1:9998", job="mediamtx"}
	0
srt_conns_bytes_received{instance="127.0.0.1:9998", job="mediamtx"}
	0
srt_conns_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}
	0
up{instance="127.0.0.1:9998", job="mediamtx"}
	1
webrtc_sessions{instance="127.0.0.1:9998", job="mediamtx"}
	0
webrtc_sessions_bytes_received{instance="127.0.0.1:9998", job="mediamtx"}
	0
webrtc_sessions_bytes_sent{instance="127.0.0.1:9998", job="mediamtx"}

You can also see some automatic metrics like up and scrape_*, and they don't have a prefix too.

[pkerney]

Thanks. Prometheus novice here.