Number of retry on conflict reschedule

[LouisLoode]

First, thanks for your work :)
Is your feature request related to a problem? Please describe.
We have a worker how create or update one of our model with external webhooks, we want to reschedule once or twice the job, not more. Actually we have some of our jobs who have infinite looping in our sidekiq.

Describe the solution you'd like
Just permitte to define max retry for lock option :while_executing and conflict strategy "reschedule", or add some documentation for it if it's already possible ?
sidekiq_options lock: :while_executing, on_conflict: :reschedule

Additional context
We try to do this but it didn't work (infinite loop).

  sidekiq_options lock: :while_executing,
                  unique_args: :unique_args,
                  on_conflict: :reschedule,
                  retry: 5

[mhenrixon]

Thanks for reporting @LouisLoode, I will consider this for the next patch release.

[axlekb]

I'm experiencing something very similar to this with:
sidekiq_options lock: :until_and_while_executing, on_conflict: :reschedule just not working at all.

When I run 5.times { MyThreeSecondJob.perform_async(my_param) }, I have ended up with 116 (!?!) new scheduled jobs that are stuck because there is a digest that seems to never get removed.

Might have something to do with the original call (5.times { ... }) ends up spiraling out of control.

2.5.3 :011 > 5.times { MyThreeSecondJob.perform_async(my_param) }
Traceback (most recent call last):
        3: from (irb):11
        2: from (irb):11:in `times'
        1: from (irb):11:in `block in irb_binding'
SystemStackError (stack level too deep)

I did test with on_conflict: :reject, and all was okay, and also lock: :while_executing, on_conflict: :reschedule and that worked as I expected.

[mhenrixon]

@axlekb partly because of a bug in removing keys. I would not use reschedule with until and while executing. I obviously need to do some more work on this one.

[robinsingh-bw]

Getting the same 'stack level too deep' bug with until_and_while_executing and on_conflict: :reschedule.

@mhenrixon Where is this bug you are referring too, I could take a look and put together a PR unless its been fixed on master already?

[mhenrixon]

@robinsingh-bw it should be fixed in master already. Thanks for the offer :)

[saiqulhaq]

should we close this issue?
since already been fixed in the recent versions

[mhenrixon]

Closing — the reschedule strategy now uses schedule_in (configurable per worker via sidekiq_options, default 5 seconds). Retry count limiting isn't built in, but you can implement a custom strategy that tracks attempts in the job hash.