Scheduling / Placement¶
Use these playbooks when the workload already declares placement or disruption intent, but the real AKS cluster shape, scheduler decisions, or drain mechanics produce an outcome you didn't expect.
flowchart TD
A[Scheduling / Placement symptoms] --> B[Spread intent not met]
A --> C[Ready capacity dips below desired]
A --> D[PDB expectations mismatch]
A --> E[Zone pools or autoscaler shape are imbalanced]Playbooks¶
| Scenario | Start Here |
|---|---|
| Pods cluster in one zone, remain Pending, or violate intended spread under capacity pressure | Topology Spread Skew Under Capacity |
| Desired replicas stay constant but Ready or Available capacity dips during failure or recovery | Ready Capacity Drops Below Desired |
| You need to confirm what a Pod Disruption Budget does and does not protect | Pod Disruption Budget Drain Contract |
| The scheduler rules are sound, but zonal node pools and autoscaler limits are shaped incorrectly | Availability-Zone-Imbalanced Node Pools and Spread Failures |