Extensions
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

best practice for alerting on ibm mq metrics(particularly oldest message age)?

tris_m
Visitor

we have some queue depth and oldest message dql queries setup in our environment and the issues is that
we're constantly getting  problem open and problem resolved notifications every time the value increases which is about every minute.
According to documentation, its best practice for dql queries interval to be set at 1 minute.z
My thinking if is the breached threshold is increasing and not decreasing, we should not be getting a resolved notification.

For example, We have ibmmq.queue.oldest_message set to alert 600 s
If a queue breaches that threshold with 700 s...and then 2 minutes later its at 750s, we should not receive a resolved because it has still breached 600.
what we get right now is a problem opening at 700(600 breached), value raises to 720, problem resolves and then immediate opens a new problem at 720.
The original problem should just stay open until it comes back down past 600.

Ideally the team wants a 24 hour window where we get one oldest message alert for a given queue and then have a day to troubleshoot it before another alert is fired. I've searched the community page for that and it seems there is moderate interest in this but there doesnt appear to be a current solution for it.

thoughts and Ideas are appreciated

0 REPLIES 0

Featured Posts