Martin Solli

To be clear, I think we want the worker to fail hard with errors like this. The exception should be logged, and the worker should stop polling for jobs. This...

I've sat down and written some tests and run through the code, and I think I've made an error in my thinking. Actually I now think the current behavior is...

Yeah, I can see that it might get you into not-so-great situations. But the alternative, to happily let clearly incorrect code continue to run, is not great either. First, I...

> > So I would ask why you've provided a false-returning ::on-polling-error? > > The worker executes jobs with side effects, so its possible for them to fail, and I...

Let me know if you have any follow-up questions regarding this. I'm closing the issue in the meantime.

Hi @binodsarkar, sorry for the late reply. The short answer is that you have the [`proletarian/on-polling-error`](https://cljdoc.org/d/msolli/proletarian/1.0.68-alpha/api/proletarian.worker) for when you want to customize the logic for when the worker should shut...

OK, and do you know how much time your handler takes to complete? Could it be that it takes longer than 90 seconds? The way the [worker works](https://github.com/msolli/proletarian/blob/main/src/proletarian/worker.clj#L53) is: 1....

Unless new information comes in, I don't think there's anything to do here. Increase the idle_in_transaction_session_timeout to at least the amount of time that the longest of your jobs are...

You can provide your own UUID-providing function with the [`:proletarian/uuid-fn` option](https://cljdoc.org/d/msolli/proletarian/1.0.89-alpha/api/proletarian.job#options). The database schema, at least for Postgres, enforces that the generated value is in fact a UUID, so that...

(One year later. Sorry! 😅) Java does not yet ship with a UUID 7 generator, but there are many libraries that can generate v7 UUIDs. Every one of them that...