Details
-
Bug
-
Resolution: Fixed
-
Critical
-
1.7.17
-
None
-
None
-
1
Description
Since the poller now runs at a higher interval, it surfaced an issue which has always been there but at higher intervals didn't really matter.
Sockets which never had a request sent in yet (maybe they are slower to connect, i.e. against the cloud) would get cleaned up too quickly, not having a chance to even take a request.