Details
-
Bug
-
Resolution: Won't Do
-
Major
-
7.6.0, 7.0.0, 7.0.1, 7.0.2, 7.0.3, 7.0.4, 7.1.4, 7.0.5, 7.1.0, 7.1.1, 7.1.2, 7.2.0, 7.1.3, 7.2.1, 7.1.5, 7.2.4, 7.0.6, 7.2.2, 7.1.6, 7.2.3, 7.2.5, 7.6.1
-
None
-
Untriaged
-
0
-
Unknown
Description
DCP Stream request was observed to run above the slow threshold of 500ms, upto ~1.1s was observed. The command executes on worker thread, and a burst of requests starved the system of executor resources.
There's a bunch of data structures/components involved the request and these are undoubtedly under contention.
Whilst we could optimise the code and aim to remove contention/slow-down, a safe first change is to move the command to a non-blocking pattern, running on NonIO - this will free executors.