Cloudwatch alarm not triggering when setpoint breached
I have the following alarm:
"Expression": "(usage_data/SERVICE_QUOTA(usage_data))*100"
Threshold: % Utilisation >= 10 for 1 datapoints within 10 minutes
Datapoints to alarm
1 out of 1
Missing data treatment
Treat missing data as missing
This does not appear to be triggering as expected. I expect the alarm to trigger if the Utilisation is greater than 10 for any datapoint in the last 10 minutes, however when it does breach the threshold, the alarm does not trigger.
- 최신
- 최다 투표
- 가장 많은 댓글
Based on what you described, it seems you should set the alarm on "SUM". If you'd use "SUM", then as total number exceeds 10 in 10 minutes duration, alarm would go "In-alarm" state and it'd trigger the target.
Also, note that if alarm is in "In-alarm" state and utilization is still greater than or equal to 10, it won't trigger the target. To re-trigger the target, it has to come back again in "OK" or "Insufficient Data" state.
Check the settings at Additional configuration -> Datapoints to alarm. This might also be a problem in your case.
Hope you find this useful.
Abhishek
I had it set on "MAX". Why would maximum not work in this case but sum would? If the maximum value exceeds 10 (which it does) then shouldn't the value returned be >10?
I'd suggest you to go through this AWS Documentation, which has very detailed explanation of evaluating an alarm. This would give you an idea, what should be data point value and evaluation period for your use case. Also look at CloudWatch statistics definitions. Hope this helps.
Check the settings at Additional configuration -> Datapoints to alarm. This might be a problem in your case.
I am using 1 of 1 datapoints to alarm. I've read the documentation but I still don't understand why this is isn't triggering.
Can you please add a snapshot of your alarm in edit mode, which would allow me to see, what's missing here. You can upload the snapshot in your answer by editing it.
Hi, can you please share the whole alarm definition? Are you alarming on average, min, max of the metric?
I am using the "MAX" of the metric
Can you fetch the activity log from this alarm? From screenshot it seems state is OK and that is valid as per graph.