phabricator.wikimedia.org should have monitoring
we want to check at least an URL from external and some running processes
we do already monitor misc-web-lb in general, but not individual services on it
we can let a check talk to nginx and check https, if we use check_http -S we get free certificate monitoring as well,
we can also talk to the Apache backend on iridium, to check if that is up, independent from misc-web
T274 is similar, but for the legalpad instance
what we already had. is the misc-web-lb up in general, on IPv4 and IPv6
https://icinga.wikimedia.org/cgi-bin/icinga/status.cgi?search_string=misc-web
suggested process monitoring of PhabricatorTaskmasterDaemon
https://gerrit.wikimedia.org/r/#/c/169585/
"currently we have around 20 processes running that match this
setting it to warn under 10 and over 40,
crit under 1 and over 50. let me know if you think i should
change these numbers
also, want me to additionally monitor processes matching these?
"PhabricatorGarbageCollectorDaemon"
"PhabricatorRepositoryPullLocalDaemon""
thresholds seem ok, a little trial and error here I guess.
also, want me to additionally monitor processes matching these?
"PhabricatorGarbageCollectorDaemon"
"PhabricatorRepositoryPullLocalDaemon""
I wouldn't as long as phd is good it controls the child procs.
commented on changeset :)
additional check for https URL from external:
https://gerrit.wikimedia.org/r/#/c/169604/
add check to LVS,hmm.. not sure if we should
https://gerrit.wikimedia.org/r/#/c/169303/
and this one, i'll likely just abandon. https://gerrit.wikimedia.org/r/#/c/169265/2
In T334250 I am wondering if we should remove the process monitoring part of this and only keep https monitoring.