We had an outage in our environment and may have been caused by a network disruption affecting Policy Server connectivity to the subnet where the User and Policy Stores are, causing performance to be much slower than usual. When checking the Policy Server logs covering the outage period, I do see the following messages:
[1362/4915][02:09:10][CServer.cpp:1431][ERROR][sm-Server-05240] 130 Stale Agent messages were discarded in the last 00:00:30.014423
[1362/2873][02:23:34][CServer.cpp:1431][ERROR][sm-Server-05240] 285 Stale Agent messages were discarded in the last 00:00:30.681736
What are these stale agent messages? What means they were discarded?
Policy Server R12.52 SP1
Web Agent R12.52 SP1
Those "Stale Agent messages" discarded are requests that the Policy Server has determined have been in the queue for too long, such that the Policy Server would not be able to reply within the Agent's RequestTimeout period. This usually happens when there is a network problem, or a huge workload, and usually is good to check the network connectivity and store logs in detail in case the problem is being caused by performance on the stores taking too much time to answer.
Usually surrounding these entries you should see other errors giving more information about what the current problem is, and why the Policy Server is taking too much to respond.