81

According to The New Apex Queueable Interface, from October 2014, we were told that chained Queueable calls would be exponentially delayed up to 64 seconds, chaining Batches would be exponentially delayed up to 240 seconds (four minutes), and suicide scheduling would be restricted to 300 seconds (five minute) minimums. This has a considerable impact on some of our code, which would would make Queueable an attractive option over Batchable (except for, of course, I don't know if we can chain queueables that callout). Today, in Spring '15, nothing behaves the way it's documented there. Are these limitations still coming, or have they been scrapped?

  • 5
    There is a critical update around Flex Queue that must be activated that I wonder if it controls this. It looks like it has an auto-activation date of May 22. Do you have that activated in your org? – Daniel Hoechst Mar 24 '15 at 20:12
  • 8
    @DanielHoechst Suicide scheduling is still running at 60 seconds, batches are running steady about 20 executes per minute per batch (it's designed to chain once per minute by estimation), so everything seems to be as it was before. I have noticed "some" slow down with batch processing, but it seems more related to performance fluctuations by system load rather than any intentional throttling (it was doing this before I turned on the Critical Update). – sfdcfox Mar 24 '15 at 20:37
  • 2
    Hey @sfdcfox what is a normal queue latency in your projects? We are seeing it as high as 10 minutes: imgur.com/x1wz3CA – dzh Jun 8 '16 at 21:59
  • 3
    FYI I've noticed in a production org recently that chaining queueable jobs is slowing down and holding queueable jobs up to a minute before processing - I'm not sure if this is new or not. I'm not seeing the same results in dev orgs because an AsyncException is still thrown if you try to chain a queueable job more than 5 times (but production orgs don't have the same limits) – Robert Watson Mar 9 '17 at 19:29
  • 2
    Queueable chain backoff limits are definitely implemented now and have been since I started working with examples of this in V39 (so it could have been many versions beforehand). My experience that significant backoff occurs after depth3 – cropredy Nov 17 '17 at 5:34

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.