I'm running a python process on an EC2 instance, and I want to alarm if it stops working properly. My plan was to have it add a custom metric data point every minute, and have a custom alarm if no data points are set for 5 minutes, say.
Is this possible? Is there another way to do what I want?
You can do that, if you use
SampleCount
for the statistic, notAverage
orSum
. Then you can trigger an alert if the number of samples is less than a certain threshold.(On a side note:
Average
orSum
would eventually go intoInsufficient Data
status rather thanAlarm
, if no samples are received at all.)