We have a cluster of 6 machines and we are monitoring the same process on each machine. We would like an alert to be generated only if 4/6, 5/6, or 6/6 components fail. How would this be possible?
The best and only way I can think of to achieve the desired results is to consolidate the criteria down to a single object/entity. This could be done by creating a SQL Query against the Orion database to return a statistical value based on the number of active failures and use this as a SQL User Experience Monitor in SAM. There you can customize the warning and critical thresholds according to the number acceptable failure conditions you can sustain.
I have never tested it - i created it on the fly, but the below trigger condition might work for your requirement , it represents 2/4 and 3/4 component failure in the same alert. Playing around with nested loops though
Rather you can go with a custom sql alert , will test it when im free and send it across to you. But then try the one i have pasted above let me know if it works or not
Hope it helps
I can't seem to get this to trigger successfully.
Triggered alerts within the Orion Advanced Alert Manager must be associated with a single object/entity. You cannot configure an alert to notify only when two or more entities match the trigger condition.
aLTeReGo thank you for kicking me on that, old school . Whats the best way to approach on it ? 4/6, 5/6, 6/6 ?
Hi stevenstadel I missed on the basics which aLTeReGo just mentioned. Guess was sleeping when I replied back to you on the post
I am sure it can be built by a custom sql query but then Im not pretty much an expert on it, the below post could be of some help to you
Advanced Alerting - Auto correlation
This post would rather talk about component being down on 2 different nodes/servers
