kube-events-action

A GitHub Action that checks Kubernetes cluster events using kube-events, and optionally posts the report as a PR comment.

Features

• Check cluster events after deployment with warning threshold
• Auto-post event reports as PR comments (updates existing comment on re-run)
• Multiple output formats: color, plain, json, markdown, table
• Group events by resource, namespace, kind, or reason
• Filter by namespace, kind, name, type, and reason
• Configurable time window with since
• Fail CI when warning count exceeds threshold

Quick Start

- name: Check cluster events
  uses: somaz94/kube-events-action@v1
  with:
    namespace: production
    type: Warning
    since: 10m

Usage

Post-deploy warning check

name: Post-deploy Check
on:
  workflow_run:
    workflows: ["Deploy"]
    types: [completed]

jobs:
  event-check:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
    steps:
      - uses: actions/checkout@v4

      - name: Setup kubeconfig
        run: echo "${{ secrets.KUBECONFIG }}" > /tmp/kubeconfig

      - name: Check events
        id: events
        uses: somaz94/kube-events-action@v1
        with:
          namespace: production
          type: Warning
          since: 10m
          threshold: '5'
          comment: 'true'
        env:
          KUBECONFIG: /tmp/kubeconfig

      - name: Fail if too many warnings
        if: steps.events.outputs.has-warnings == 'true'
        run: |
          echo "Warning count: ${{ steps.events.outputs.warning-count }}"
          exit 1

Filter by kind and reason

- uses: somaz94/kube-events-action@v1
  with:
    namespace: production
    kind: Pod
    reason: BackOff,Unhealthy,Failed
    since: 5m
    output: table
    comment: 'false'

All namespaces summary

- uses: somaz94/kube-events-action@v1
  with:
    all-namespaces: 'true'
    type: Warning
    since: 1h
    summary-only: 'true'

Group events by namespace

- uses: somaz94/kube-events-action@v1
  with:
    all-namespaces: 'true'
    group-by: namespace
    type: Warning
    since: 10m

Group events by reason

- uses: somaz94/kube-events-action@v1
  with:
    namespace: production
    group-by: reason
    since: 10m
    output: table
    comment: 'false'

JSON output for downstream processing

- name: Get event report
  id: events
  uses: somaz94/kube-events-action@v1
  with:
    namespace: production
    output: json
    comment: 'false'

- name: Process result
  run: echo '${{ steps.events.outputs.result }}' | jq '.summary.warningCount'

Scheduled cluster health audit

name: Cluster Health Audit
on:
  schedule:
    - cron: '0 8 * * *'

jobs:
  audit:
    runs-on: ubuntu-latest
    steps:
      - name: Check warnings
        id: events
        uses: somaz94/kube-events-action@v1
        with:
          all-namespaces: 'true'
          type: Warning
          since: 24h
          output: json
          comment: 'false'

      - name: Notify Slack
        if: steps.events.outputs.has-warnings == 'true'
        run: |
          curl -X POST "${{ secrets.SLACK_WEBHOOK }}" \
            -d "{\"text\": \"Cluster warnings: ${{ steps.events.outputs.warning-count }} in last 24h\"}"

Inputs

Input	Description	Required	Default
namespace	Filter by namespace (comma-separated)	No	all
kind	Filter by involved object kind (comma-separated)	No	all
name	Filter by involved object name (comma-separated)	No	all
type	Filter by event type: Normal, Warning	No	all
reason	Filter by event reason (comma-separated)	No	all
since	Show events newer than duration (e.g., 5m, 1h)	No	10m
group-by	Group events by: resource, namespace, kind, reason	No	resource
output	Output format: color, plain, json, markdown, table	No	markdown
summary-only	Show summary statistics only	No	false
all-namespaces	Show events from all namespaces	No	false
threshold	Fail if warning count exceeds this value (0 = do not fail)	No	0
comment	Post result as PR comment	No	true
version	kube-events version to install	No	latest
token	GitHub token for PR comments and authenticating the version lookup (raises the GitHub API rate limit)	No	${{ github.token }}

Outputs

Output	Description
result	Full event report output text
warning-count	Number of warning events detected
has-warnings	true if warnings were detected, false otherwise

Threshold Behavior

The action does not fail by default when warnings are detected. Set threshold to control failure behavior:

# Fail if more than 5 warnings
- uses: somaz94/kube-events-action@v1
  with:
    threshold: '5'

# Never fail (default)
- uses: somaz94/kube-events-action@v1
  with:
    threshold: '0'

Or use has-warnings output to control your workflow:

- name: Fail on any warning
  if: steps.events.outputs.has-warnings == 'true'
  run: exit 1

Combined Usage with kube-diff-action

Use kube-diff-action to detect manifest drift and kube-events-action to check cluster warnings — all in one workflow.

Post-deploy validation

name: Deploy & Validate
on:
  push:
    branches: [main]

jobs:
  deploy-and-validate:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
    steps:
      - uses: actions/checkout@v4

      - name: Check drift before deploy
        id: diff
        uses: somaz94/kube-diff-action@v1
        with:
          source: file
          path: ./manifests/
          namespace: production
          output: markdown
          comment: 'false'

      - name: Deploy manifests
        if: steps.diff.outputs.has-changes == 'true'
        run: kubectl apply -f ./manifests/

      - name: Check cluster events after deploy
        id: events
        uses: somaz94/kube-events-action@v1
        with:
          namespace: production
          type: Warning
          since: 5m
          threshold: '3'
          comment: 'true'

      - name: Summary
        run: |
          echo "Drift detected: ${{ steps.diff.outputs.has-changes }}"
          echo "Warnings after deploy: ${{ steps.events.outputs.warning-count }}"

Scheduled cluster health check

name: Cluster Health Check
on:
  schedule:
    - cron: '0 */6 * * *'

jobs:
  health-check:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Check manifest drift
        id: diff
        uses: somaz94/kube-diff-action@v1
        with:
          source: kustomize
          path: ./overlays/production/
          output: json
          comment: 'false'

      - name: Check cluster warnings
        id: events
        uses: somaz94/kube-events-action@v1
        with:
          all-namespaces: 'true'
          type: Warning
          since: 6h
          output: json
          comment: 'false'

      - name: Notify on issues
        if: steps.diff.outputs.has-changes == 'true' || steps.events.outputs.has-warnings == 'true'
        run: |
          curl -X POST "${{ secrets.SLACK_WEBHOOK }}" \
            -d "{\"text\": \"🔍 Cluster issues detected\nDrift: ${{ steps.diff.outputs.has-changes }}\nWarnings: ${{ steps.events.outputs.warning-count }}\"}"

License

This project is licensed under the Apache License 2.0 — see the LICENSE file for details.