Manual retry API

[ppawlucz]

It would be nice if pgqueuer had an API for manual retry of a failed job, after all retries were exhausted. What do you think?

[janbjorge]

I’m open to adding something like this, but I think it’s worth first understanding the concrete use-case. Right now PGQueuer doesn’t log the payload or headers - only metadata - so a true “retry the same job” flow would require additional work, either by extending the log schema or by having the application provide the payload again.

I’m also curious why the existing retry mechanisms aren’t sufficient in your case. PGQueuer already supports automatic retries (including backoff at the executor level and recovery of stalled jobs), so understanding what scenario requires a manual retry after those are exhausted would help shape a clean API.

[ppawlucz]

I want a job to execute only once and let the client decide if he wants to retry it or not if it failed. I could enqueue new job with the same payload but that would mean I would also need to update/copy state of that job that I hold in another table with the new job id

[janbjorge]

Could you take a look at; #507 and give feedback?