web-platform-tests / pulls.web-platform-tests.org

[Deprecated] Some functionalities are now provided by wpt-pr-bot https://github.com/web-platform-tests/wpt-pr-bot
7 stars 23 forks source link

Improve monitoring of pulls website #43

Open mattl opened 6 years ago

mattl commented 6 years ago

Restart things when they're down, etc.

foolip commented 6 years ago

@mdittmer already has some monitoring at https://bit.ly/ecosystem-infra-status, but we removed ecosystem-infra from notifications because of noise.

mattl commented 6 years ago

Oh yeah, and I have my own monitoring too, but it just tells me when its down. I'd like to improve that to actually do something.

foolip commented 6 years ago

Oh, so responding with 504 counts as being up? I have something very rudimentary in https://github.com/whatwg/misc-server/blob/master/test/wiki.js, but I guess you'd want something that runs more than every 24 hours :)

mattl commented 6 years ago

No, it doesn't count as being up. Ideally, when that happens something can restart the service after a few retries. This helps in the situation where things are down at 3am more than when things are down at 3pm on a weekday.