You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This behavior relies on the update being aborted, and retried (our intention is that these retries are done by gRPC interceptors and/or history client in the Frontend service). However, currently, the error responses are being returned to the SDK caller; the desired behavior occurs because the SDK client retries the ResourceExhaustedFailure. So:
We need to make the server retry this internally (at least, that was out original intention)
I think that we want to confirm that retries will be spread out in time such that races with the CAN handling are very unlikely.
The text was updated successfully, but these errors were encountered:
temporalio/features:features/continue_as_new/updates_do_not_block_continue_as_new is a feature test that confirms that, when an update is admitted while a WFT is in-flight, and that WFT CANs, that the update lands on the next run.
This behavior relies on the update being aborted, and retried (our intention is that these retries are done by gRPC interceptors and/or history client in the Frontend service). However, currently, the error responses are being returned to the SDK caller; the desired behavior occurs because the SDK client retries the
ResourceExhaustedFailure
. So:The text was updated successfully, but these errors were encountered: