Build, test, release—this might work for a simple app or a startup. But in regulated enterprises? Things get a lot more complicated. You’re not just shipping code; you’re juggling a maze of processes and expectations. Think about automatically generating a Software Bill of Materials (SBOM), navigating complex change processes with human approvals, or implementing branching strategies that track every ticket across multiple environments. And you still need to keep your pipelines evolvable and find errors fast.

The 'pipeline as code' approach is a natural fit for these challenges. To keep your pipelines scalable and evolvable, you need a solid architecture that can handle edge cases and unexpected complexities.

In this session, I’ll share a decade of experience building enterprise CI pipelines across platforms like Gitlab CI/CD and Azure DevOps. From designing branching strategies that ensure traceability across multiple environments to handling compliance-driven change processes with human approvals, I’ll walk you through real-world challenges I’ve faced. You’ll hear about the mistakes I made—like overmodularizing pipelines or neglecting useful debug output—and the lessons I learned to build pipelines that are scalable, adaptable, and easy to debug.

## Take Home Ideas

- Evolvability: Pipelines are never set in stone. Especially in enterprise environments you often get new requirements from regulations. Or you have to support a deployment to a whole new environment in short time. Choosing a modular approach combined with parameterization can help you do without messing up your pipeline code.
- Debuggability: Debugging pipelines is a pain as you often just have to run them to see what happens. Useful debug output is often neglected but is a cornerstone of moving fast when developing pipelines.

## Additional Material

- Former Talks: https://fiveandahalfstars.ninja/talks/johannesdienst_talks.html

## Notes

This talk is a collection of all the learnings I had over the years building enterprise pipelines with Gitlab CI/CD, GitHub Actions and Azure DevOps.