Scheduled Scaling - Autopilot by Xano

The Scheduled Scaler overrides HPA minReplicas during defined time windows — pre-warming pods before predictable traffic bursts. CRD: ScheduledScalePolicy (namespaced)

Example

apiVersion: ops-ai.io/v1alpha1
kind: ScheduledScalePolicy
metadata:
  name: morning-burst
  namespace: xano-tenant
spec:
  hpaName: backend-hpa
  mode: active
  windows:
    - name: morning-login-surge
      minReplicas: 50
      preScaleMinutes: 5
      scaleUpSpeed: Instant
      disableScaleDown: true
      schedule:
        type: Weekly
        startTime: "08:30"
        endTime: "10:30"
        timezone: "America/Los_Angeles"
        daysOfWeek: [Monday, Tuesday, Wednesday, Thursday, Friday]

Each window can raise minReplicas, pre-scale ahead of the window (preScaleMinutes), and optionally override the HPA’s scaling behavior for the duration of the window. Original values are saved and restored when the window deactivates.

For the full set of HPA behavior overrides a window can apply — scale-up/down speed presets, disableScaleDown, CPU target — see HPA Behavior Overrides.

Dashboard

The Scheduled Scales page shows all policies with active-window status and activation history — so you can confirm a window actually fired.