Closed isaacseymour closed 9 years ago
The cleanest approach is probably for people to disable retrying globally, but if they can't do so to use this approach - worth adding a note / clarification to that effect?
I'm not sure, I kinda prefer doing it per-job. Doing in globally requires putting something like ActiveJob::QueueAdapters::SidekiqAdapter::JobWrapper.sidekiq_options(retry: false)
, which feels prety fragile, and I'm not sure it's even possible with Que (you can just set the retry delay to many seconds).
Yeah, downside is that you lose any ability to manually retry failed jobs when you do it this way.
I'm :+1: on this either way - it's an improvement, and people should know their own queueing back-ends.
Yeah, at this point it all gets super backend-specific, so mostly just want to make it really clear that it's something to be considered.
Added another para about the alternative (Sidekiq-specific since there's no dead/failed job queue in Que, and don't know about others).
:+1:
@greysteil I think this complements #28 . The same caveats that apply to Sidekiq also apply to Que, which we already support. Does this look good to you?