Back to all scenarios
Scenario #437
Scaling & Load
Kubernetes v1.22, AWS EKS

Disrupted Service During Pod Autoscaling in StatefulSet

Pod autoscaling in a StatefulSet led to disrupted service due to the stateful nature of the application.

Find this helpful?
What Happened

Scaling actions impacted the stateful application, causing data integrity issues.

Diagnosis Steps
  • 1Reviewed StatefulSet logs and found missing data after scale-ups.
  • 2Found that scaling interfered with pod affinity, causing service disruption.
Root Cause

StatefulSet’s inherent behavior combined with pod autoscaling led to resource conflicts.

Fix/Workaround
• Disabled autoscaling for stateful pods and adjusted configuration for better handling of stateful workloads.
Lessons Learned

StatefulSets need special consideration when scaling.

How to Avoid
  • 1Avoid autoscaling for stateful workloads unless fully tested and adjusted.
Previous ScenarioNext Scenario