You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Deployment tools in general (kubectl, spinnaker, argo, etc.) use this condition to signal that a deployment failed. However, FIAAS only checks for the desired vs ready pods which is an incomplete view of what k8s is surfacing.
This issue is to discuss wether FIAAS needs to follow this same pattern and consider the progress condition in deployments.
Some possibilities could be:
include a timeout in the ApplicationSpec to create the deployment with a given timeout (default to 0, so it's barckwards compatible with current behaviour).
ignore the condition as FIAAS does today and rely exclusively on the desired vs ready pods as of today.
others?
The text was updated successfully, but these errors were encountered:
Kubernetes has the notion of https://kubernetes.io/docs/concepts/workloads/controllers/deployment/#progress-deadline-seconds to understand when a deployment is making progress (or not) via a condition. Unless specified, that timeout is set to 10m.
Deployment tools in general (kubectl, spinnaker, argo, etc.) use this condition to signal that a deployment failed. However, FIAAS only checks for the desired vs ready pods which is an incomplete view of what k8s is surfacing.
This issue is to discuss wether FIAAS needs to follow this same pattern and consider the progress condition in deployments.
Some possibilities could be:
The text was updated successfully, but these errors were encountered: