[
https://issues.apache.org/jira/browse/FINERACT-2965?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Adam Saghy closed FINERACT-2965.
--------------------------------
Resolution: Not A Problem
> Add optional jitter to the exponential backoff of command retries
> -----------------------------------------------------------------
>
> Key: FINERACT-2965
> URL: https://issues.apache.org/jira/browse/FINERACT-2965
> Project: Apache Fineract
> Issue Type: Bug
> Reporter: Jenny Liu
> Priority: Major
> Fix For: Unknown
>
>
> Retries of a command that failed on an optimistic lock or another concurrency
> failure wait 1 s, then 2 s, with no randomization. Callers that failed
> together therefore retry together and collide again. Add an optional
> randomization factor
> fineract.retry.instances.executeCommand.exponential-backoff-jitter (default
> 0, behavior unchanged). Measured on a running instance with 20 concurrent
> deposits to one savings account: without jitter the retries cluster and 9 to
> 14 of 20 requests still fail with HTTP 409; with jitter 0.5 the waits are
> spread and 20 of 20 succeed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
