Seeing issues with resource versions

[ericuldall]

The following error is popping up in logs and may be causing some unknown issues:

I0622 12:00:00.015384       1 scaling-controller.go:118] FAILED TO UPDATE SCHEDULED SCALER STATUS: [HPA_NAME] - Operation cannot be fulfilled on scheduledscalers.scaling.k8s.restdev.com "[HPA_NAME]": the object has been modified; please apply your changes to the latest version and try again

Needs further investigation.

[ericuldall]

Just ran a couple tests in my GKE server and the updates appear to be working even though the error is thrown. I'll keep an eye on this for now and see about getting the errors resolved when I have some time.

[m0rganic]

we see this error quite a bit .. what happens as a result is very bad -- the schedule just doesn't kick in and we miss critical windows where production services are not scaled to the appropriate capacity levels.. we've taken a look at our processes and have not been able to track down what causes this in all cases

Could the issue be due to the fairly outdated version of the kube client package?

[ericuldall]

It's possible. Perhaps newer changes to the api have created an adverse side effect that wasn't present on previous tests. The package has not been updated for latest versions of kubes either.

[scr-oath]

Does #29 fix this?

[guongle-ssense]

@m0rganic @ericuldall have you both resolved this as I'm observing this issue in EKS.