Page MenuHomePhabricator

Diamond log level set to DEBUG spams syslog
Closed, ResolvedPublic

Description

Despite configuration it looks like the new version of diamond logs as DEBUG level with --log-stdout --foreground, spamming syslog

Event Timeline

Migrating my comment from T171638, as it may be useful for debugging:

I believe it is on stretch, because I have not seen it on other hosts, but it could be because a newer package or other things:

Jun 22 10:10:46 db2072 python[20568]: 1498126246.73        [CPUCollector:21046:DEBUG]        Collection took 3 ms
Jun 22 10:10:50 db2072 python[20568]: 1498126250.08        [DiskUsageCollector:20948:DEBUG]        Collection took 6 ms
Jun 22 10:10:51 db2072 python[20568]: 1498126251.08        [NetworkCollector:20869:DEBUG]        Collection took 14 ms

While it is not a huge problem, it puts a lot of garbage on the syslog (making more difficult to debug, and probably not useful at all unless they are errors). Specially because I saw it in some cases (servers) complain about special devices not being able to show available disk space (so that is 20 lines every minute):

Jul 12 11:17:15 dbstore2002 python[777]: 1499858235.21        [NetworkCollector:1124:DEBUG]        Collection took 19 ms
Jul 12 11:17:15 dbstore2002 python[777]: 1499858235.21        [NetworkCollector:1124:DEBUG]        Collection took 19 ms
Jul 12 11:17:15 dbstore2002 python[777]: 1499858235.21        [NetworkCollector:1124:DEBUG]        Collection took 19 ms
Jul 12 11:17:16 dbstore2002 python[777]: 1499858236.15        [CPUCollector:1387:DEBUG]        Collection took 3 ms
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys since it is of typ
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /proc since it is of ty
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /dev since it is of typ
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /dev/pts since it is of
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /run since it is of typ
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/kernel/security si
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.85        [DiskSpaceCollector:1448:DEBUG]        Ignoring /dev/shm since it is of
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /run/lock since it is o
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup since it
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/systemd 
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/pstore since it
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/perf_eve
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/cpu,cpua
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/memory s
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/freezer 
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/net_cls,
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/pids sin
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/devices 
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/cpuset s
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/fs/cgroup/blkio si
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /sys/kernel/debug since
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /dev/hugepages since it
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /proc/sys/fs/binfmt_mis
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /dev/mqueue since it is
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Ignoring /proc/sys/fs/binfmt_mis
Jul 12 11:17:17 dbstore2002 python[777]: 1499858237.86        [DiskSpaceCollector:1448:DEBUG]        Collection took 7 ms
Jul 12 11:17:18 dbstore2002 python[777]: 1499858238.99        [LoadAverageCollector:1420:DEBUG]        Collection took 3 ms
Jul 12 11:17:22 dbstore2002 python[777]: 1499858242.74        [DiskUsageCollector:1451:DEBUG]        Collection took 9 ms
Jul 12 11:17:27 dbstore2002 prometheus-mysqld-exporter[790]: time="2017-07-12T11:17:27Z" level=error msg="Error pinging mysqld: dial 
Jul 12 11:17:33 dbstore2002 python[777]: 1499858253.52        [TCPCollector:1476:DEBUG]        Collection took 18 ms

The collectors on stretch already had to be tuned before, so my thesis would be more config that is different from jessie: https://gerrit.wikimedia.org/r/361878

faidon raised the priority of this task from Medium to High.Jul 27 2017, 1:22 PM
faidon moved this task from Inbox to In progress on the observability board.

Change 368177 had a related patch set uploaded (by Filippo Giunchedi; owner: Filippo Giunchedi):
[operations/puppet@production] diamond: ship systemd override file

https://gerrit.wikimedia.org/r/368177

Change 368177 merged by Filippo Giunchedi:
[operations/puppet@production] diamond: ship systemd override file

https://gerrit.wikimedia.org/r/368177

Logging level has been set to INFO again for diamond, following up on parent task