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
{{ message }}
This repository has been archived by the owner on Oct 24, 2023. It is now read-only.
Users expect their critical system pods to always be scheduled as needed regardless of any actions they take, like tainting nodes for workload specific scheduling.
Today because AKS and AKS-engine system pods are controlled by the addon manager and the default pod spec enforced by it does not use any tolerations or priority this leads to the pods being evicted after for example a taint and/or not be scheduled in the tainted nodes anymore.
Proposed Solution
Add PriorityClasses and wildcard tolerations to the critical system pods
The text was updated successfully, but these errors were encountered:
Problem Summary:
Users expect their critical system pods to always be scheduled as needed regardless of any actions they take, like tainting nodes for workload specific scheduling.
Today because AKS and AKS-engine system pods are controlled by the addon manager and the default pod spec enforced by it does not use any tolerations or priority this leads to the pods being evicted after for example a taint and/or not be scheduled in the tainted nodes anymore.
Proposed Solution
Add PriorityClasses and wildcard tolerations to the critical system pods
The text was updated successfully, but these errors were encountered: