KEDA CRON Scaling

• Select a timezone from the IANA Time Zone Database.
• Define cron expressions for when scaling should start and when it should end.
• Set the exact replica count (desiredReplicas) for the scheduled window.

​

Why use Cron scaling?

​

How the Cron trigger works

• timezone — an IANA time zone string (for example, Asia/Kolkata). Use values from the IANA Time Zone Database.
• start and end — cron expressions that define when the scheduled replica count should be applied and when it should stop being enforced.
• desiredReplicas — the number of replicas to maintain during the scheduled period (KEDA expects this as a string).

​

Cron field reference

​

Example trigger metadata

triggers:
- type: cron
  metadata:
    # Required
    timezone: Asia/Kolkata        # Value from the IANA Time Zone Database
    start: "0 6 * * *"           # At 06:00 (6:00 AM)
    end: "0 20 * * *"            # At 20:00 (8:00 PM)
    desiredReplicas: "10"        # Note: value must be a string

​

Example ScaledObject

• minReplicaCount: 0 to allow scaling down to zero outside scheduled windows.
• cooldownPeriod (seconds) — the time KEDA waits after the last active trigger before scaling down. This helps prevent rapid down-scaling after brief spikes.

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: example-cron-scaledobject
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app-deployment
  minReplicaCount: 0
  cooldownPeriod: 300          # 300 seconds (5 minutes) before scaling down after inactivity
  triggers:
  - type: cron
    metadata:
      timezone: Asia/Kolkata
      start: "0 6 * * *"
      end: "0 20 * * *"
      desiredReplicas: "10"

​

Notes on cooldown and behavior

• cooldownPeriod is the number of seconds KEDA waits after the last active trigger observation before attempting to scale down. For long scheduled windows (e.g., 06:00–20:00) this setting usually does not impact the scheduled period, but it can help avoid oscillation when other scalers are active.
• When the end cron expression matches, KEDA removes the scheduled desired replica count and other scalers (or the default behavior) determine the replica count. That may trigger a gradual scale-down according to your configuration.

​

When to use KEDA Cron scaler

• Predictable, time-based traffic (daily business hours, scheduled promotions, batch processing windows).
• Need to guarantee minimum capacity for a known window (pre-warming).
• Desire to scale to zero outside scheduled windows to save resources.

​

References

• KEDA Cron scaler: https://keda.sh/docs/latest/scalers/cron/
• IANA Time Zone Database: https://www.iana.org/time-zones
• Kubernetes Deployments: https://kubernetes.io/docs/concepts/workloads/controllers/deployment/