Rethink how we monitor CI failures on our main branches
Originally created by @CyrilBrulebois on #17624 (Redmine)
Scope
-
short term: how do we handle this with our current tooling? - Mostly solved by https://gitlab.tails.boum.org/tails/ci/-/blob/master/scripts/tails-notmuch-jenkins for boyska and intrigeri
-
mid term: are there cheap tooling improvements that would help? - Existing tooling: https://gitlab.tails.boum.org/tails/ci/-/tree/master/scripts
-
long term: what would our tools do ideally? how do we get closer to this? - Related to whether we switch to GitLab CI for builds & tests of Tails images
Initial report
As the RM responsible for 4 releases in a row, I’m not comfortable with having to keep an eye on Jenkins failures right afterwards; but given the increased tempo on the Firefox side, it doesn’t leave much room to breathe anyway.
This led me to wonder why Jenkins access/notifications should be limited to RMs; extending that at least to FT would allow for some members to keep an eye on those as well, without having to have one RM triage issues first.
This is a reminder to start a discussion on this topic, likely across -rm & -foundations.
For background, some initial discussion on this topic happened in 2019, in the “Analyzing failure notifications for Jenkins jobs” thread, on the RMs mailing list.