That is exactly why I think a process might be useful. I dont believe we have a clear way to answer that question of "What is the current test status? What are current known issues? And how are we making progress towards no flakyness?"
At least not that I know of. One of us can spend some time looking around recent test failures and answer that with some confidence. But as this seems like a repeat pattern, having a running list of known issues would be useful. What do you think? Are you suggesting we not have the process of rotation but just do one off analysis? On Tue, Oct 4, 2016 at 1:47 PM, Li Li <l...@cloudera.com> wrote: > It is better if we can get some idea about the current test status, like > how many known or unknown flaky tests, which tests are disabled and with > what reason, any existing jiras tracking these failure tests. > > Also does it include both the master and feature branch? > > On Tue, Oct 4, 2016 at 1:28 PM, Sravya Tirukkovalur <sra...@apache.org> > wrote: > > > Hi all, > > > > As the project grows, I think it might be useful to have a process and > > rotation of responsibility on how we maintain the health of test runs > which > > is a crucial piece in project health. > > > > I documented a draft of the process, please provide your inputs if you > > think it would be useful also please tell if you think it would not be > > useful :-) > > > > https://docs.google.com/document/d/1rvxMvONo- > > ABqW4w4NyCDXz6eMrLFB4GJexMFSE0DG9w/edit?usp=sharing > > > > Would love to hear if any one has experience with similar processes in > > other communities? > > > > Regards, > > Sravya > > > -- Sravya Tirukkovalur
