Vanished hosts are not deleted for Kubernetes DCD

Troubleshooting
,
1

CMK version: 2.1.0p20
OS version: Rocky Linux 8.4

So I have configured Kubernetes monitoring using the steps described in the documentation.
Everything works great except for pods that don’t exist anymore are still available as host in CheckMK and reporting errors like [piggyback] Missing data WARN, Got no information from host CRIT but also check_icmp: No hosts to check.

I have triple checked and toggled the option Automatically delete hosts without piggyback data is enabled in the Dynamic Host Connection.

Did I missed some hidden configuration option somewhere or is this a bug in CheckMK?

The hosts that give those errors should have be deleted 4 days ago.

2

Hi,

this also seems to be this bug

I think p21 will be released shortly.

2 Likes
3

Hi!
Unfortunately it seems is still present on 2.1.0p33. But this time things are a bit different
My OP departement made a “re-deploy” though every POD changed his name.
I expected to see “bulk” deletion of the POD with “old” name, indeed they are still there.
Is there any step that I can do to go deep with “debug”?