This message was deleted.

To illustrate what I have in mind:

😍 , I really like this. I think recently CircleCI went into a similar direction and we found some of the metrics really actionable. In terms of metrics: build duration (especially over time to see whether you are regressing and need to reinvest) amount of flaky builds (as that should always be low so people trust the suite and results) merge on red (as a big warning sign) maybe also avg. time per contributor (if a rarely touched project runs 1h it doesn't really matter, but if your main app runs 1h in CI something is probably wrong) I would also find some alerts like: your CI suite runtime increased 30% last week, this is likely the PR where it started, super useful.

Thanks for the input @Johannes Würbach ! Some very helpful points, especially correlation with team size as well as merge on red. For the latter I wonder, is that a warning on the build or on the PR … or both?

I was more thinking of a weekly report or some kind of summary page, but I guess on the PR / build would also work if you can detect it reliably enough. One slow build could often be an outlier, but if the 95% (or some other number) changes across multiple builds that is probably a strong indicator of permanent slow down, which can only be improved with actively changing something again.

A bit like this @Johannes Würbach? 🙂

Yes and no. I think a summary dashboard isn't really actionable here. My first question would be: What is the worst performing project across my org so I can do a targeted investment to fix this.