I have a use case to alert in case of SLA miss. My application emits metric on startTime (M1) & endTime (M2). If my job completes, I will be able to know SLA misses by doing metric math like (M2-M1) and having alerting on this.
But if my job is stuck, I still want to get alerted by computing (currentTime-M1) (may be on scheduled basis). Is this possible with AWS CloudWatch? Non-AWS based approaches & solutions are also welcome!!