Pod Disruption Budget (PDB)

Keep minimum pods available during node drains and cluster upgrades

📋Configuration Files

pdb.yml

PDB for Deployment — at least 1 pod available during voluntary disruptions

yaml

1apiVersion: policy/v1
2kind: PodDisruptionBudget
3metadata:
4  name: myapp-pdb
5  namespace: apps
6spec:
7  minAvailable: 1
8  selector:
9    matchLabels:
10      app: myapp
11---
12# Alternative: maxUnavailable for large fleets
13# spec:
14#   maxUnavailable: 1
15#   selector:
16#     matchLabels:
17#       app: myapp

deployment-with-labels.yml

Deployment labels must match PDB selector

yaml

1apiVersion: apps/v1
2kind: Deployment
3metadata:
4  name: myapp
5  namespace: apps
6spec:
7  replicas: 3
8  selector:
9    matchLabels:
10      app: myapp
11  template:
12    metadata:
13      labels:
14        app: myapp
15    spec:
16      containers:
17        - name: app
18          image: myapp:latest
19          ports:
20            - containerPort: 8080
21          readinessProbe:
22            httpGet:
23              path: /health
24              port: 8080
25            initialDelaySeconds: 10
26          livenessProbe:
27            httpGet:
28              path: /health
29              port: 8080
30            initialDelaySeconds: 30

📄

Step 01

Apply PDB

(01)Apply manifests

Linux

1kubectl apply -f deployment-with-labels.yml
2kubectl apply -f pdb.yml
3kubectl get pdb -n apps
4kubectl describe pdb myapp-pdb -n apps

(02)Test drain respects PDB

Linux

1# voluntary disruption — should respect minAvailable
2kubectl drain <node> --ignore-daemonsets --delete-emptydir-data
3# undo: kubectl uncordon <node>