As of release 3.5.0 it will be possible to configure checks so that the non-existence of instances does not trigger an alarm. The default configuration won’t change, meaning that an UNKNOWN will still be sent when the performance getter (performance collectors) don’t find volumes, disks, luns, etc… The behavior for each performance collector can be changed with the switch ‑‑no_instances=OK|WARNING|… In addition, there is the possibility to change the behavior of checks with ‑‑no_store=OK|WARNING|… and ‑‑no_instances=OK|WARNING|… The reasoning behind this, as described in this article , is that it makes sense to set up the monitoring configuration of some objects, such as LUNs, independently from their availability, since they are often created and deleted dynamically. This configuration can then be used without having to make continuous adjustments.
Related Posts
- No instances should be ok?! Mar 9, 2016
- -timeout Bug Mar 8, 2016
- DiskCount counts SSD spare disks Mar 8, 2016
- Monitoring Lun Sizes Jan 21, 2016
- Monitoring Cluster Peer Health Jan 8, 2016
Comments