[CheckMK] monitoring should not alert on PB-size mounted disk space

[acozine]

Expected behavior

A mounted disk that has hundreds of GB of space free should not alert just because the total disk used is above 80%. For example, /mnt/diglibdata1/pas is alerting because 1.30 PiB are used of 1.40 PiB. True, that is 93% usage, but 0.10 PiB == 100,000 GB. We have plenty of space.

Actual behavior

CheckMK returns a critical error.

Steps to replicate

Go to staging CheckMK and look at the slavery-staging1 host entry.

Impact of this bug

We get pure-noise alerts and may ignore alerts we should be paying attention to.

Implementation notes, if any

We solved this for basic disk space by creating a rule called Adjustable filesystem warnings. However, the rule is either incorrectly configured or does not apply to CIFS mounted disks. We are seeing the same critical alerts on the bibdata-staging machines, which also mount drives from diglibdata1.