[LabelBot] Alarming #30
Description
With the last deployment, we noticed that we didn't really know where to look for while investigating an issue. In the end, it was the SQS queue that was not being consumed. Afterwards, we had exceptions being thrown #23 that were not bubbled up.
It would be good if there were some monitors/alarms that could be attached that trigger if the SQS queue gets too full (flood), messages don't get consumed (outage) or if there is an exception during execution.