Closed cretz closed 1 month ago
Timeouts at least seem to be working properly: https://github.com/temporalio/sdk-core/commit/e64b05b8f26f132deba7c32eab7c682eeff861d0
I can confirm that timeout does work as expected on update start. I am unsure why/how I came to the conclusion it didn't. Changing title/description to reflect this as a long-poll timeout issue only.
Describe the bug
Two issues:
set_timeout
is used by lang on non-long-poll call (retry client or not), it seems to get overridden by the 30s default timeout despite the interceptor checking for thegrpc-timeout
header first.~ EDIT: Not able to replicate, not an issuepoll_workflow_task_queue
andpoll_activity_task_queue
are the only calls that get the 70s long poll timeout today. This means things like waiting for workflow result get the default 30s which actually returns empty response in 20s. We need to also set that long poll timeout forupdate_workflow_execution
,poll_workflow_execution_update
andget_workflow_execution_history
(whenwait_new_event
is true).