Do You Trust Pod Disruption Budgets?

KubeCon + CloudNativeCon Europe 2026

Session Date: March 24, 2026 — London

What Are Pod Disruption Budgets?

  • Kubernetes resource that limits voluntary disruptions to pods
  • Controls how many pods can be down simultaneously during operations
  • Critical for maintaining availability during node drains, upgrades, and scaling

Why Trust Matters

  • PDBs are only respected during voluntary disruptions (drains, upgrades)
  • Involuntary disruptions (node crashes, OOM kills) bypass PDBs entirely
  • Karpenter, cluster autoscaler, and node upgrade tools rely on PDB compliance
  • Misconfigured PDBs can block node drains indefinitely

Best Practices

Practice Description
Set maxUnavailable Prefer over minAvailable for flexibility
Avoid 0 maxUnavailable Can permanently block node operations
Test PDBs Simulate drains to verify behavior
Monitor PDB status Alert on PDBs blocking operations
Combine with topology spread Reduce blast radius across zones

Integration with Karpenter

  • Karpenter respects PDBs when consolidating and draining nodes
  • Karpenter Node Disruption Budget adds node-level controls
  • Together they enable safe, automated node lifecycle management
  • Spot interruptions (2-min warning) work within PDB constraints

Key Takeaways

  1. PDBs are essential but only protect against voluntary disruptions
  2. Misconfigured PDBs are a common source of blocked operations
  3. Combine PDBs with topology spread constraints for real availability
  4. Karpenter's Node Disruption Budget complements pod-level PDBs
  5. Test your PDBs — don't just deploy and hope

Questions?

KubeCon EU 2026 — London