Issue Summary
Asynchronous webhooks were using Jira’s request cache outside of the appropriate context. This caused the webhook executor to generate the JSON payload without access to some caches.
This problem was detailed in JRASERVER-78768, but the fix for the problem was incomplete. As a result, even on Jira versions with that fix, webhooks triggered by worklog operations, issue linking, some administrative actions were still not using cache.
Steps to Reproduce
- Configure one or more [worklog event] webhooks, including the JSON payload.
- Induce a high-load (100+ updates/seconds) of matching events.
Expected Results
The webhook executor can process each webhook and build its payload body in a reasonable amount of time. Database pressure is reasonable.
Actual Results
Call stacks that should otherwise quickly return with cached values are delayed.
Workaround
Jira 9
- Increase webhook executor thread pool and queue sizes, as described on
JRASERVER-74492. - Switch back to synchronous webhooks.
Jira 10
Increase webhook executor thread pool and queue sizes, as described on Tuning webhook performance parameters in Jira Data Center.
- is cloned from
-
JRASERVER-78768 Async webhooks do not use request cache in Jira Data Center (Invalid use of RequestCache by thread: webhook-dispatcher)
-
- Closed
-
- is related to
-
JRASERVER-79070 Webhook payloads generated for all webhooks cause performance issues
-
- Closed
-
- details
-
WILDT-338 Loading...