Skip to main content

5 Continuous Delivery & Deployment

Overview​

Continuous Delivery (CD) and Continuous Deployment (CD) are the second half of the CI/CD pipeline, following Continuous Integration (CI).
They focus on automatically delivering and deploying tested and validated code into production or production-like environments.

  • Continuous Delivery: Ensures that every code change is always in a deployable state. Deployments to production may still require manual approval.
  • Continuous Deployment: Extends delivery by automating deployments into production without human intervention, making software updates available to users immediately.

Objectives of CD Deployment​

  • Automated, repeatable deployments across environments.
  • Rapid and reliable release cycles without downtime.
  • Consistent infrastructure to reduce deployment errors.
  • Improved release confidence through staging and pre-production checks.
  • Faster feedback from end-users.

Key Stages of a CD Deployment Workflow​

1. Artifact Retrieval​

  • Retrieves build artifacts from the artifact repository (e.g., Nexus, JFrog Artifactory, Docker Registry).
  • Ensures deployment is based on tested and versioned artifacts, not raw code.

2. Infrastructure Provisioning​

  • Automates the creation of servers, containers, or cloud resources using Infrastructure as Code (IaC).
  • Tools: Terraform, AWS CloudFormation, Ansible, Pulumi.
  • Ensures environments (dev, staging, production) are consistent and reproducible.

3. Configuration Management​

  • Ensures that applications are deployed with correct configurations (environment variables, secrets, endpoints).
  • Tools: Ansible, Chef, Puppet, Kubernetes ConfigMaps & Secrets, Vault.

4. Deployment Strategies​

Different strategies ensure zero downtime and safe releases:

  • Recreate Deployment

    • Old version is stopped before the new one is started.
    • Simple but causes downtime.

  • Rolling Updates

    • Gradually replaces old instances with new ones.
    • Minimizes downtime.

  • Blue-Green Deployment

    • Two environments (Blue = current, Green = new).
    • Switch traffic to the new version instantly once validated.
    • Easy rollback.

  • Canary Deployment

    • Release new version to a small percentage of users first.
    • Monitor results before rolling out to everyone.

5. Automated Testing in Deployment​

  • Smoke Tests: Quick checks after deployment.
  • End-to-End Tests: Validate real-world workflows.
  • Performance & Load Tests: Ensure stability under traffic.
  • Security Tests: Validate vulnerabilities in production-like environments.

6. Monitoring & Observability​

  • Continuous monitoring ensures smooth deployments and quick detection of issues.
  • Metrics: CPU, memory, response times, error rates.
  • Logs: Aggregated and analyzed for troubleshooting.
  • Tools: Prometheus, Grafana, ELK Stack, Datadog, New Relic, OpenTelemetry.

7. Rollback Mechanisms​

  • Rollback strategies are critical in case of failed deployments.
  • Options:
    • Revert to last stable version from artifact repository.
    • Traffic switch in blue-green or canary deployments.
    • Automated rollback triggered by monitoring alerts.

Example CD Deployment Workflow​

ASCII Diagram​

       +--------------------+
       |  Artifact Storage  |
       +--------------------+
                │
                ▼
   +------------------------+
   | Infrastructure as Code |
   +------------------------+
                │
                ▼
   +------------------------+
   | Configuration & Secrets|
   +------------------------+
                │
                ▼
   +------------------------+
   |   Deployment Strategy  |
   | (Blue-Green, Canary)   |
   +------------------------+
                │
                ▼
   +------------------------+
   | Automated Validation   |
   +------------------------+
                │
                ▼
   +------------------------+
   | Monitoring & Rollback  |
   +------------------------+