Use perform_now instead of perform #41

New Issue

2025-12-29T00:21:44+01:00

adam commented

2025-12-29 00:21:44 +01:00

Originally created by @NeilBetham on GitHub (Feb 24, 2017).

Would it be possible to use ActiveJob's peform_now instead of perform? The main reason I ask is that using peform_now will call all of the ActiveJob callbacks and error handlers. The error handler callback is particularly important for reporting problems to a error reporting system like Sentry. Is there a reason perform was used instead of perform_now?

Originally created by @NeilBetham on GitHub (Feb 24, 2017). Would it be possible to use ActiveJob's `peform_now` instead of `perform`? The main reason I ask is that using `peform_now` will call all of the ActiveJob callbacks and error handlers. The error handler callback is particularly important for reporting problems to a error reporting system like Sentry. Is there a reason `perform` was used instead of `perform_now`?

adam commented

2025-12-29 00:21:45 +01:00

@MohamedBassem commented on GitHub (May 2, 2017):

Yes please. I spent 2 hours trying to find why job failures doesn't trigger airbrake notifications until I found that the rescue_from of the ActiveJob is not triggered because Job.new.perform is used instead of perform_now.

@MohamedBassem commented on GitHub (May 2, 2017): Yes please. I spent 2 hours trying to find why job failures doesn't trigger airbrake notifications until I found that the `rescue_from` of the ActiveJob is not triggered because `Job.new.perform` is used instead of `perform_now`.

adam commented

2025-12-29 00:21:45 +01:00

@edwardmp commented on GitHub (Aug 8, 2017):

I agree. Is there any reason perform is used because of perform_now?

@edwardmp commented on GitHub (Aug 8, 2017): I agree. Is there any reason perform is used because of perform_now?

adam commented

2025-12-29 00:21:47 +01:00

@edwardmp commented on GitHub (Aug 8, 2017):

As a workaround you can do something like this:

class NewRegistrationProcessor
  def perform
    NewRegistrationProcessorJob.perform_now
  end
end

Crono.perform(NewRegistrationProcessor).every 1.minute

But I agree it would be better if perform_now is called by default

Edit: in my case, this still doesn't report the error to Rollbar. It seems Crono catches the error elsewhere: https://github.com/plashchynski/crono/blob/5b72e08222047ee6454fd3110904336386e995b9/lib/crono/job.rb#L81

so you would need to monkey patch the handle_job_fail method to get this to work.

@edwardmp commented on GitHub (Aug 8, 2017): As a workaround you can do something like this: ``` class NewRegistrationProcessor def perform NewRegistrationProcessorJob.perform_now end end Crono.perform(NewRegistrationProcessor).every 1.minute ``` But I agree it would be better if perform_now is called by default Edit: in my case, this still doesn't report the error to Rollbar. It seems Crono catches the error elsewhere: https://github.com/plashchynski/crono/blob/5b72e08222047ee6454fd3110904336386e995b9/lib/crono/job.rb#L81 so you would need to monkey patch the `handle_job_fail` method to get this to work.

adam commented

2025-12-29 00:21:47 +01:00

@sina-s commented on GitHub (Sep 8, 2018):

I have the exact same use-case, reporting errors to Sentry. I came up with another work around that doesn't need monkey patching Crono. Assuming your jobs are subclass of ApplicationJob, define your perform inside ApplicationJob and change the perform method name to something else (in my case perform!) in your job class. Not sure how graceful this work around is, as it requires to change all of your job classes.

class ApplicationJob < ActiveJob::Base
  def perform
    perform!
  rescue StandardError => e
    Raven.capture_exception(e)
  end 
end

class MyFrequntTask < ApplicationJob
  def perform!
     ...
   end
end

@sina-s commented on GitHub (Sep 8, 2018): I have the exact same use-case, reporting errors to Sentry. I came up with another work around that doesn't need monkey patching Crono. Assuming your jobs are subclass of `ApplicationJob`, define your `perform` inside `ApplicationJob` and change the `perform` method name to something else (in my case `perform!`) in your job class. Not sure how graceful this work around is, as it requires to change all of your job classes. ``` class ApplicationJob < ActiveJob::Base def perform perform! rescue StandardError => e Raven.capture_exception(e) end end ``` ``` class MyFrequntTask < ApplicationJob def perform! ... end end ```

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: starred/crono#41