Slow networking from misc.lizard for check_mirrors_cronjob.sh
Since 2022-05-13, every day, 1 run of this cronjob fails a little bit after 05:00 UTC with a timeout, after 300m. I see unusually slow transfer rates from 10+ mirrors for each of these runs (when the whole thing does not timeout even before having tried so many mirrors), which makes me suspect a problem on our side, or with the colo connectivity, rather than suddenly tons of our mirrors becoming slow.
This prevents the cronjob from checking the other mirrors, which may leave us unknowingly with broken mirrors, which can break initial installation and upgrade UX, hence P:Elevated.
However, I don't think we should spend more than 1-2h on this before giving up: instead I think we should work on tails#18263 (closed).
Edited by intrigeri