I have created a Quality Gate mostly focussed on what has been happening ‘lately’ - but one metric is about the overall current coverage in the project. The Quality Gate is set as the default - and all projects use it. The Leak Period is set to previous_version.
I have several projects that (correctly) fail the quality gate - and fail it only due to low coverage, but I have several projects that are below the coverage minimum, but show as Passed for the quality gate.
Thanks for the response Ann. Yes, I did check that - and have just validated again. The three projects I mentioned about all use the same “Default” quality gate - which is my custom one.
Two of them use the default Python Quality Profile, and one uses that as well as the default C# Quality Profile. But - the quality profile should be independent of the Quality Gate Coverage metric - correct?
Thanks for reporting this. I’ve reproduced it in 6.7.5, but not in 7.4 (altho I can’t find the ticket related to the fix). It appears that the Warning condition is evaluated first, and evaluation stops if it’s true. We’re not likely to tackle this in an LTS Bug fix. The workaround would be to remove the Warning condition.
Note that we’re planning to drop the ability to set Warning conditions in the “near” future:
Thanks Ann. I may have actually found an issue on my side for a couple of the builds. I had missed the fact that a few of them had not built in quite a while - and I had only added the “Coverage” metric a few weeks ago (seems like forever ago). Rerunning these seemed to make most of them clear up.
As you say - I may still have an issue with the warnings… but I can work around that as you suggest!