I'm making the switch from Jenkins to Harness at my organization and I'd love to hear any tips or experiences from those who have done the same. I've heard some concerning things about Harness, specifically regarding its compatibility with Terraform. I've been told that build artifacts don't persist during a build run, and that to keep artifacts afterward, they need to be copied to S3. I really hope that information is incorrect! Do people find it difficult to work with Terraform in Harness because it's still relatively new, or is it intentional to push users further into the Harness ecosystem? I'm looking for guidance or warnings about this transition, especially regarding the artifact persistence and needing to deal with a lot of YAML configuration instead of the declarative pipeline style I'm used to in Jenkins. By the way, do you recommend learning Harness through their free courses or diving straight into the documentation? What works best for understanding the platform?
3 Answers
The best way to learn Harness is really through hands-on experience. I created a personal account and used the free tier to play around with it. I even set up a homelab to deploy and understand its behaviors—having the documentation open while experimenting helped a ton!
We use Harness for Infrastructure as Code with Terraform, and it’s been smooth sailing for us! We developed custom Harness delegates to execute our IAC code without issues. Just keep in mind that your experience might vary depending on your setup.
You definitely want the artifact persistence issues to be true! Keeping artifacts linked with your build can complicate things. It's much better to have a clean separation between the lifecycle of your builds and your artifacts. The key to succeeding with tools like Harness is to fully embrace the change and let go of any Jenkins habits. Keep an open mind and explore Harness's capabilities!
I think it's more about adapting to new strategies and tools, which does feel overwhelming. You're right that it's not just about Harness; it's a total shift in how we work!
But what about artifacts related to test results or Terraform plans generated during the build? Wouldn't separating those create gaps?