Understanding XYZ Team DevOps Pipeline

In this guide, we’ll walk through the GitLab CI/CD workflow for building, testing, and deploying the XYZ team’s Node.js application. Every commit in the GitLab repository triggers a sequence of automated stages—running tests, measuring code coverage, building Docker images, and deploying to Kubernetes clusters—culminating in manual approval and production rollout.

Pipeline Overview

Stage	Description	Failure Behavior
1. Unit Testing	Install dependencies and run unit tests	Stops on failure
2. Code Coverage	Parallel jobs: test-report & coverage-analysis	Coverage errors ignored
3. Containerization	Build, smoke-test, and push Docker image	Stops on failure
4. Deploy to Development	Apply Kubernetes manifests to the dev cluster and fetch ingress URL	Stops on failure
5. Integration Testing	Verify service health via the dev endpoint	Stops on failure
6. Manual Approval	Human gate before production deployment	Pauses pipeline until approved
7. Deploy to Production	Apply manifests to the prod cluster and fetch ingress URL	Stops on failure
8. Post-Deployment Integration Testing	Validate production health via the live endpoint	Stops on failure

1. Unit Testing

Install dependencies with npm and execute all unit tests. Any failure aborts the pipeline immediately.

# Install dependencies
npm install

# Run unit tests
npm test

# (Optional) Archive or download test reports

Note

Store your test reports (e.g., JUnit XML) as artifacts for later analysis or reporting.

2. Code Coverage

Run two parallel jobs:

Test Report: Fails on test errors.
Coverage Analysis: Runs coverage but ignores errors to avoid blocking downstream stages.

# Job A: Test Report
npm install
npm test
# Job B: Coverage Analysis (errors ignored)
npm install
npm run coverage
# Archive coverage-report artifact

Note

Ignoring coverage errors ensures that a slight dip in metrics doesn’t halt deployments.

3. Containerization

Build and push a Docker image tagged with the commit SHA. Optionally, smoke-test the container before pushing.

# Build image
docker build -t myregistry/xyz-app:${CI_COMMIT_SHA} .

# Optional smoke test
docker run --rm myregistry/xyz-app:${CI_COMMIT_SHA} npm test

# Push to registry
docker push myregistry/xyz-app:${CI_COMMIT_SHA}

4. Deploy to Development

Apply Kubernetes manifests to the development cluster and retrieve the ingress hostname.

# Apply manifests to dev context
kubectl apply -f k8s/ --context dev

# Fetch Dev ingress URL
INGRESS_URL=$(kubectl get ingress xyz-app-ingress \
  --context dev \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')

echo "Dev URL: https://${INGRESS_URL}"

Note

Ensure your kubeconfig is configured with the dev context. See Kubernetes docs.

5. Integration Testing

Validate the development deployment by hitting the health endpoint.

curl --fail https://${INGRESS_URL}/live

6. Manual Approval

A developer or reviewer must approve the pipeline before production rollout. If declined, the pipeline stops here.

Warning

Do not skip this step. Manual gates help prevent unintended production changes.

7. Deploy to Production

Once approved, deploy the same manifests to the production cluster.

kubectl apply -f k8s/ --context prod

INGRESS_URL=$(kubectl get ingress xyz-app-ingress \
  --context prod \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')

echo "Prod URL: https://${INGRESS_URL}"

8. Post-Deployment Integration Testing

Confirm that the production service is live and healthy.

curl --fail https://${INGRESS_URL}/live

Commands Summary

# 1. Unit Testing
npm install
npm test

# 2. Code Coverage (parallel jobs)
npm install && npm test          # Test Report
npm install && npm run coverage  # Coverage Analysis (ignored errors)

# 3. Containerization
docker build -t myregistry/xyz-app:${CI_COMMIT_SHA} .
docker run --rm myregistry/xyz-app:${CI_COMMIT_SHA} npm test
docker push myregistry/xyz-app:${CI_COMMIT_SHA}

# 4. Deploy to Dev
kubectl apply -f k8s/ --context dev
INGRESS_URL=$(kubectl get ingress xyz-app-ingress \
  --context dev \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')
curl --fail https://${INGRESS_URL}/live

# 6. Deploy to Prod
kubectl apply -f k8s/ --context prod
INGRESS_URL=$(kubectl get ingress xyz-app-ingress \
  --context prod \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}')
curl --fail https://${INGRESS_URL}/live

Links and References

Watch Video

Watch video content