rucio / probes

Common Nagios probes to monitor Rucio
Apache License 2.0
0 stars 23 forks source link

Retire check_unlocked_replicas #140

Open dchristidis opened 1 month ago

dchristidis commented 1 month ago

ATLAS wishes to retire the check_unlocked_replicas probe.

Its first query counts the number of replicas with a set tombstone across all RSEs. In an instance where having unlocked replicas is the norm, this number conveys no useful information. The per-RSE information is already calculated in check_deletable_replicas and stored as RSE usage counter.

Its second query further limits the first to replicas whose tombstone is more than two hours in the past. Since the tombstone is set to either the time of the creation of the replica or its last use, this query makes no sense.

@ericvaandering, if CMS would like to keep using it, then it will be moved to the cms directory. Otherwise, it will be moved to the attic directory.

ericvaandering commented 1 month ago

We do use this, so let's move it. Can we do this with the Rucio 36 release?

dchristidis commented 1 month ago

I’m genuinely curious, how do you extract any piece of useful information out of these two numbers?