CI Failing due to CPU limits on Action Runners [bug]

[Adam-D-Lewis]

Github hosted action runners have 2 cores. With the latest PR adding Prometheus/Grafana to Qhub by default, the CI is failing because we are requesting ~2.5 cores from our minikube cluster, but as stated, we only have 2 cores.
#733

I want to try using CIRun (https://cirun.io/), as an alternative which allows us to spin up Action Runners of various sizes only as needed. @aktech do we have a Quansight account with access to CIRun?

[aktech]

the CI is failing because we are requesting ~2.5 cores from our minikube cluster

Ah interesting.

@aktech do we have a Quansight account with access to CIRun?

Yep we do have one via @quansight-bot

[Adam-D-Lewis]

Amit let me know that some work has been done on this at https://github.com/aktech/qhub/tree/run-tests-cirun

[Adam-D-Lewis]

Also, from Amit

Also, when you do that, I would suggest to move the kubernetes tests job to a separate file:
https://github.com/Quansight/qhub/blob/528bb850158d78d2eabdb561e1d37040052119a1/.github/workflows/test.yaml#L93

[Adam-D-Lewis]

I'm waiting for a Quansight org owner to add the Cirun github app to the qhub repo

[aktech]

This was fixed here: 265d6c5