Create a /metrics endpoint for Prometheus model

[dridk]

Like many web server applications, it would be great to have a default /metrics endpoint exposing metrics for Prometheus data model. Metrics must be pull by Prometheus. The dagster Prometheus plugin work only in push , which is not recommended.

Data exposed examples;

• asset count
• assets status
• run queued
• run running
  ...

Example:

Minio metrics
https://min.io/docs/minio/kubernetes/upstream/operations/monitoring/metrics-and-alerts.html

Starrocks
https://docs.starrocks.io/docs/administration/management/monitoring/Monitor_and_Alert/

I can work on it if you guide me.
What is the best approach to create an endpoint? GraphQL ?
What metrics do you need ?

[danielgafni]

It might be the case that the daemon, the web server, and the code locations might require separate dedicated endpoints