🧪 Still using Classic Pipelines? Time to switch to YAML!
Even though Microsoft has been promoting YAML Pipelines as the standard in Azure DevOps for quite some time, I’m still surprised to see Classic Pipelines being used. While most teams have adopted YAML for new pipelines, there’s often that one release pipeline still running as a Classic Pipeline.
These pipelines were often set up years ago and “work just fine.” And that’s exactly the issue: many teams stick with Classic Pipelines simply because they still work.
But technology doesn’t stand still. As the name suggests, Classic Pipelines are now considered outdated. YAML Pipelines offer more flexibility, better reusability, and integrate better with modern DevOps practices.
🤔 Why do companies still use Classic Pipelines?
At most companies, YAML is the default for new pipelines. However, I still come across Classic Pipelines that are actively being used. There are usually three main reasons for this:
The pipeline still works. “If it works, why change it?” This is a common mindset among teams who set up pipelines once and never look back.
The team is used to Classic Pipelines. Developers prefer to focus on building features. Rewriting a Classic Pipeline into YAML takes time and effort, which many teams prefer to spend on new features.
There’s a misconception that YAML doesn’t support everything. A common misunderstanding is that approvals or manual interventions are only possible in Classic Pipelines. This is not true. YAML provides comprehensive options for approvals and manual interventions — often with more flexibility than Classic Pipelines.
🚀 Switching to YAML? It took me some time to adjust too
I know firsthand that switching to YAML Pipelines can feel like a big change. When I first started using YAML, I had to get used to writing pipelines as code. The visual interface of Classic Pipelines felt familiar, and YAML seemed complicated at first.
For a long time, I found myself quickly creating pipelines in the Classic UI because it was easier.
But now, I use YAML daily — not just in Azure DevOps, but also in GitHub Actions, GitLab, Docker, and Kubernetes. I wouldn’t go back to Classic Pipelines.
And that’s yet another reason for me to switch to YAML. For most teams, YAML Pipelines have already become the standard. When I encounter issues with a Classic Pipeline, it often takes me more time to figure out what might be causing the error. As a result, I usually end up rewriting the pipeline to YAML right away.
🌟 What are the benefits of YAML Pipelines?
Here are the key benefits of switching to YAML Pipelines:
🗂️ 1. Pipelines in Version Control
YAML Pipelines are stored in your repository alongside your code. This means you can track, review, test, and manage pipeline changes through pull requests, giving you more control over your CI/CD processes.
While Classic Pipelines also offer a version history, it’s much harder to track and compare changes. It’s often difficult to figure out when or why a particular change was made.
🔄 2. Reusable templates
With YAML, you can create reusable templates that can be shared across projects. This makes it easier to set up new pipelines. Templates ensure consistency and save teams time by reducing manual pipeline configuration.
It’s similar to using Task Groups in Classic Pipelines — but more powerful and scalable across projects. Templates can be created at various levels, such as for tasks, jobs, stages, or entire pipelines.
⚙️ 3. Automating complex workflows
YAML supports complex workflows with conditions, stages, and multi-job pipelines. These workflows can be fully automated without relying on the UI.
For example, you can easily run different tests and builds in parallel, depending on the branch or environment you’re working in.
🔐 4. Better integration with modern DevOps practices
In DevOps, we talk a lot about Infrastructure as Code (IaC), Continuous Integration (CI), and Continuous Deployment (CD). YAML Pipelines fit perfectly into this philosophy by enabling Pipeline as Code.
🛠️ 5. Fewer maintenance issues and easier troubleshooting
When your team primarily uses YAML Pipelines but still has a few remaining Classic Pipelines, troubleshooting issues in those Classic Pipelines can become increasingly challenging.
As knowledge and experience within the team shift toward YAML, fewer people will remember how Classic Pipelines work. If an issue arises in a Classic Pipeline over time, resolving it will likely take longer, and you may ultimately have to decide whether to fix it or migrate to YAML. That’s not a decision you want to make when a critical bug fix needs to be deployed to production as soon as possible.
📋 How to migrate from Classic Pipelines to YAML
Here are a few steps to help you migrate from Classic Pipelines to YAML:
Start with new projects. For new pipelines, use YAML from the start to avoid future migrations. I already see this happening at most organizations.
Disable the creation of new Classic Pipelines. Azure DevOps allows you to disable the creation of Classic Pipelines at the project or organization level. This prevents teams from creating new Classic Pipelines while still allowing existing ones to run. (No more pouring water into a leaky bucket!)
Migrate existing Classic Pipelines step by step. Start with the simplest pipelines and build from there. This way, you can quickly achieve success and build confidence within the team.
Use templates. If you convert a pipeline to YAML that you might reuse, turn it into a template for other projects to use.
Leverage tools like Yamlizr. Yamlizr is a useful tool for converting Classic Pipelines to YAML automatically. While it’s not a magic solution, it can save a lot of manual work and give you a solid starting point. Always review the generated YAML to ensure everything is correctly converted.
🔧 Next blog: A deeper dive into migration steps
In my next blog, Ill provide more detailed examples and tips to make your migration to YAML Pipelines smoother.
💡 Need help optimizing your CI/CD Pipelines?
Migrating from Classic Pipelines to YAML Pipelines is an investment that will future-proof your pipelines. But I understand it’s not always easy to free up time for this.
Looking for someone to help guide you through the process?