I built Build Change Investigator because one of the annoying parts of Jenkins failures is that the error itself is usually obvious, but figuring out what changed since the last successful build can take a while.
The plugin automatically compares a failed/unstable build with the last successful one and pulls together:
- commits since the last good build
- changed files
- build/revision metadata
- a reduced failure-log excerpt
- the evidence in one place so you can narrow down the likely regression faster
The core investigation works without AI. There’s also optional AI-assisted analysis if you want it, with support for OpenAI, Anthropic Claude, AWS Bedrock, Azure OpenAI, Gemini, Ollama, and OpenAI-compatible endpoints.
No Jenkinsfile changes are required. Once installed, the Build Change Investigation action appears on applicable failed/unstable builds.
It’s live in the Jenkins plugin ecosystem:
https://plugins.jenkins.io/build-change-investigator/
If anyone tries it on a real pipeline, I’d genuinely like feedback on whether the change correlation is useful and what you’d want it to show differently.
[link] [留言]