Richard KorantengDatabase Engineering
Case study

Self-service CI/CD pipeline for database deployments with Liquibase & Azure DevOps

How an Oracle team moved from manual releases to gated, automated Liquibase pipelines in Azure DevOps.

2 min read
Azure DevOps continuous deployment pipeline for database releases
Challenge

Database changes moved through manual processes, isolated from CI/CD, so releases were slow and risky.

Solution

A self-service Azure DevOps pipeline that validates, previews, and promotes changes with Liquibase.

Results

Changes reach Dev in minutes, production releases are gated by approvals, and every change is traceable.

Stack

Oracle · Liquibase · Azure DevOps

Problem

I recently worked with a customer who struggled to promote database changes to production quickly and safely. Their changes often remained locked in manual processes, isolated from modern CI/CD practices.

Impact

This gap caused:

  • Delays in delivering new features
  • Increased risk of human error in production deployments
  • A disconnect between dev and ops teams

Solution

Here's a better way - something secure, automated, and developer-friendly. I developed an Azure DevOps CI/CD pipeline for database deployments that:

  • Validates database changes using Liquibase
  • Automates deployments to Dev, Staging, and Production
  • Enforces approval gates and rollback previews
  • Promotes changes safely to production
  • Requires no DBA intervention for routine changes

Validate and package once, then promote the same artifact

CI pipeline: validate and package

  1. Validate changelog
  2. Show pending changes
  3. Generate SQL preview
  4. Generate rollback SQL
  5. Deploy to Dev
  6. Tag database
  7. Generate build metadata
  8. Build artifacts
  9. Publish artifact
    • changelog.xml
    • /scripts/*.sql
    • update-preview.sql
    • rollback-preview.sql
    • version.txt
    • liquibase.properties (optional)

CD pipeline: test and promote to production

  1. Triggered by CI
  2. Downloads CI artifact
  3. Deploys to staging and production using the changelog
  4. References version and tag for auditing
CI publishes one versioned artifact; CD deploys that same artifact to staging and production.

Outcome

Implementing this solution allows:

  • Developers to commit database changes and see them deployed to Dev within minutes
  • Promotion to Production is fully automated but gated for approvals
  • Dev and Ops teams have visibility and traceability of every change
  • DBAs are no longer bottlenecks - we're enablers of innovation
  • The business can respond faster to customer needs and market changes

Summary

By automating database deployments through a self-service CI/CD pipeline, developers are empowered to safely deliver changes without waiting on manual DBA intervention. This accelerates release cycles, reduces downtime, and improves team productivity, leading to faster time-to-market.