Details
-
Bug Report
-
Resolution: Won't Do
-
L3 - Default
-
None
-
7.12.0
-
None
Description
I've been trying to configure reliable at-least-once delivery (see https://www.futurice.com/blog/at-least-once-delivery-for-serverless-camunda-workflow-automation)
I notices the retry parameter has some stranger behavior. I want two things
- Infinite retries
- A sane backoff policy
According to your docs "The configuration follows the ISO_8601 standard for repeating time intervals. In the example, R5/PT5M means that the maximum number of retries is 5 (R5) and the delay of retry is 5 minutes (PT5M)."
So for my first attempt I used R/PT30S for an infinite retry with a 30 second wait between reties. THIS DOES NOT WORK. For some reason the engine just retries 3 times. This is not ISO 8061 compliant. OK, so I set it to R99999/PT30S and things work as expected. But then the 30 second delay kills throughput so I try to add a backoff policy with:
R1/PT1S,R1/PT3S,R99999/PT30S
THIS DOES NOT WORK! It goes back to retrying only 3 times!
All in all, its very confusing and it takes me a very long time to see the misconfiguration manifest in production. And I still cannot figure out how to achieve sane backoff with very high retries. I see two bugs, R/ doesn't do infinite retries. Comma separated ISO 8601 don't seem to count retries up properly.
mgm-controller-panel
This is the controller panel for Smart Panels app
Attachments
Issue Links
- is related to
-
CAM-11759 Add support for timer cycle definitions for failedJobRetryTimeCycle
-
- Closed
-