I have a python UDF task which was working for a few days before it hit an exception. To summarise the task looks at the “x-forwarded-for” headers and does a geo-lookup on the address contained within. However when we received two addresses which is a valid header, the code didn’t account for that scenario and failed, which took a couple of days to notice.
So this has raised two questions:
- Is it possible to receive alerts on a Kapacitor task which enters the “Executing: false” state?
- Can Kapacitor automatically restart a failed task?
We have updated our UDF to account for the problem scenario but it would be nice to catch any similar situations sooner.