Hi DOJO,
The default alarm storm protection configuration seems to be the following:
-
If there are more than 2000 alarms for the same parameter name, Cube automatically enables alarm storm protection by grouping all these alarms.
Once the number of alarms for that parameter drops below 1900, the protection is deactivated. -
If, within a 10-second time window, more than 250 alarms are received, Cube will also enable alarm storm protection by delaying new alarms for 5 minutes.
In this case, the protection will be disabled once fewer than 100 alarms enter Cube within a 10-second window.
I’d like to confirm whether these default settings are recommended regardless of the type of system, or if there are guidelines to calculate more appropriate values based on the number of agents in the cluster, elements, etc.
Thanks in advance!
Hi Catarina,
These defaults are OK for most systems. As far as I know, there are no guidelines to calculate adjusted settings.
If a system has so many alarms that the alarm storm protection happens too often, you can consider these techniques to reduce the alarms:
- review the alarm templates, so no unneeded alarms are generated
- for alarms that toggle a lot, add hysteresis
- use filtered alarm tabs to reduce the total number of alarms displayed at once