When shutting down, restarting, or deploying a new version of an app, all processes have 30 seconds to finish, including jobs. If a worker is stuck in a job that takes more than that, the process will receive a SIGKILL. Note that this is not the only reason why we want to do the heavy lifting in batches, locks and memory consumption are other examples.
On Wednesday, September 5, 2018 at 11:58:53 AM UTC+2, Greg Navis wrote: > > I'm a bit confused with your remark about the 30 second grace period. > There's a 30 second timeout on web dynos but worker dynos can work up to 24 > hours + jitter (that's when Heroku's mandatory dyno cycling kicks in). > Could you clarify what you mean by that? > > You may run into issues if you run the Heroku Scheduler because before > launching the job it terminates its previous instance if it's still working > (so if you have a job that runs every 10 minutes you have 10 minutes to > finish and so on). > > Best regards > Greg Navis > -- You received this message because you are subscribed to the Google Groups "Ruby on Rails: Core" group. To unsubscribe from this group and stop receiving emails from it, send an email to rubyonrails-core+unsubscr...@googlegroups.com. To post to this group, send email to rubyonrails-core@googlegroups.com. Visit this group at https://groups.google.com/group/rubyonrails-core. For more options, visit https://groups.google.com/d/optout.