This morning graphite1003 (the machine hosting cassandra metrics) ran out of disk space. I've added some from the VG but there isn't a whole lot left in the VG itself (66G).
I'm proposing to regularly clean up metric files that haven't been modified say in six months.
Description
Description
Details
Details
Subject | Repo | Branch | Lines +/- | |
---|---|---|---|---|
graphite: cleanup cassandra metrics | operations/puppet | production | +6 -0 |
Related Objects
Related Objects
- Mentioned Here
- T173436: Delete graphite metrics for old CFs
Event Timeline
Comment Actions
Change 386587 had a related patch set uploaded (by Filippo Giunchedi; owner: Filippo Giunchedi):
[operations/puppet@production] graphite: cleanup cassandra metrics
Comment Actions
Also, there are metrics that are no longer being emitted and can be cleaned up, cf T173436: Delete graphite metrics for old CFs.
Comment Actions
Change 386587 merged by Filippo Giunchedi:
[operations/puppet@production] graphite: cleanup cassandra metrics
Comment Actions
Mentioned in SAL (#wikimedia-operations) [2017-10-26T13:32:44Z] <godog> force delete old graphite cassandra metrics - T179057