I saw similar output, and the cause was a backup job hogging all the disk I/O. Pollers would take a very long time to run because
the virtual server itself was waiting to access the disk. (Was running the server on a RAID 5 using 7K drives. Bumped them up to 10K SAS, adjusted the poller count and haven’t had an issue since)
You may find this link useful.
https://jira.observium.org/browse/OBS-1729
Brady Farver
From: observium [mailto:observium-bounces@observium.org]
On Behalf Of Victor Marinho via observium
Sent: Wednesday, April 24, 2019 07:30
To: observium@observium.org
Cc: Victor Marinho <victor@softhousenet.com.br>
Subject: [Observium] problems Graphs
Good morning, the observium graphics are getting this way what can it be?
--
Victor Marinho
Soft House Net
Fone: 34 3241-6536 // 3241-3400
Celular: 34 98813-9966