[🚀 Feature]: Prometheus Integration for Grid4 Components

[aviw19]

Description

I see there are no metrics being pushed by critical components such as router/distributor. Only source of observability is traces or the graphQL endpoint. But to track use-cases such as increased errors on the /session endpoint, latencies etc, pushing of metrics seems to be essential.

Would like to also know if there's something already existing and can be leveraged in a distributed world.

Have you considered any alternatives or workarounds?

No response

[selenium-ci]

@aviw19, thank you for creating this issue. We will troubleshoot it as soon as we can.

Selenium Triage Team: remember to follow the Triage Guide

[pujagani]

Hey! We have traces, and ideally, they should be able to provide a pattern for failing sessions, including how long it took to complete a new session request, etc. Depending on where these traces are exported to. I do understand that metrics have a different purpose, but we currently do not have anything instrumented. We are open to receiving contributions though!

[aviw19]

Thanks for the turnaround Puja.

1. Can we export metrics out of traces? If not, it's pretty hard to apply them to a production distributed world, since we'll have to collect and then re-process them.
2. Would the OSS community be open to a overall Prometheus proposal? I can identify points where we could integrate this.

[diemol]

@aviw19 why can't you use a transformer to feed information into Prometeus?