Gardener 1.30 Maintenance‑Window Blocks Node‑Pool Scaling During Active Periods

Diagnosing Deployment Failures During Maintenance Windows

When a Gardener 1.30 Shoot enters an active maintenance window, the controller emits a MaintenanceWindowEntered event and subsequently blocks any node‑pool scaling actions. A scaling request made during this period is logged as Upgrade blocked by maintenance window and is queued until the window closes. This behavior surfaces as a failed deployment when the desired replica count is not applied in time.

Operators often rely on automated scaling or manual patching of node‑pool replicas. The current documentation does not specify whether an override is possible or how to safely perform critical scaling during a maintenance window, leaving uncertainty about safe deployment practices.

Given this ambiguity, how should operators plan scaling during scheduled maintenance windows? What mechanisms, if any, can be used to temporarily override the block for urgent scaling? Should future releases introduce a flag to allow critical scaling while a maintenance window is active?