I am running onWatch in a k8s cluster where I already have accessible / easy to leverage monitoring via Prometheus/Grafana/Alert-Manager.
Exposing Prometheus format metrics would address two needs:
- A way to alert when my Codex auth token has expired... this seems to happen often and I don't notice it until sometimes hours/days after it has expired and I suddenly find that i'm near my limits.
- A way to combine the token usage data with other data that already resides in prometheus so I can graph my codex/co-pilot token usage right along side other metrics like those from my github-actions exporter
I am running onWatch in a k8s cluster where I already have accessible / easy to leverage monitoring via Prometheus/Grafana/Alert-Manager.
Exposing Prometheus format metrics would address two needs: