Bits of .NET

🛡️ Don’t Drain All Pods at Once

Node draining, cluster upgrades kill pods. PodDisruptionBudget (PDB) ensures minimum availability during voluntary disruptions.

📝 PDB Examples

# At least 2 pods available at all times
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: app-pdb
spec:
  minAvailable: 2
  selector:
    matchLabels:
      app: myapp

# No more than 1 pod unavailable
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: app-pdb
spec:
  maxUnavailable: 1
  selector:
    matchLabels:
      app: myapp

# For single-replica apps (minimum 1)
spec:
  minAvailable: 1

⚙️ Check PDB Status

# List PDBs
kubectl get pdb

# Describe PDB
kubectl describe pdb app-pdb

# Output shows:
# Allowed disruptions: 2
# Current disruptions: 0
# Desired healthy: 3
# Current healthy: 3

# Simulate disruption (will be blocked if PDB violated)
kubectl drain node-1 --ignore-daemonsets

💡 Use Cases

• Databases: Ensure primary replica never disrupted
• API gateways: Keep at least 2 pods running
• Queue workers: Prevent all consumers from stopping
• Leader election: Keep at least 1 pod (leader) online