On two of our Ubuntus 20.04 (test/prod) we got repeatable (at least once per day, often two times) one of this alerts:
monit alert -- Filesystem flags changed XXX
- filesystem flags changed to ro,relatime
monit alert -- Does not exist XXX
- unable to read filesystem '/dev/XXX' state
Directly with the next check (2mins) this got resolved. And we got the corresponding message.
My monit configuration on both systems is:
check filesystem XXX with path /dev/XXX if space usage > 80% for 5 times within 15 cycles then alert if inode usage > 50% then alert if read rate > 1 MB/s for 5 cycles then alert if read rate > 500 operations/s for 5 cycles then alert if write rate > 1 MB/s for 5 cycles then alert if write rate > 500 operations/s for 5 cycles then alert if service time > 10 milliseconds for 3 times within 5 cycles then alert
We already have this problem since 5.27.0. Also with 5.27.1. And still with 5.28.0.
We already checked with our server administrator, but it looks like a monitoring problem by monit, because everything else is working. And it’s resolved automatically directly by the next check.
Do you observed similar behaviors?
The problem is not major, but I try to teach our team to look at this alerts if I’m not available, and they are getting confused, if there were alerts.