These questions are specific to the DotNet agent.
1) In the event of a network outage and the agent can’t communicate to its collector,
a) Where does the agent store metrics while the connection is down?
b) How long are metrics stored? i.e., is there a time limit or size limit before older metrics are overwritten with new ones?
c) When the connection is restored are all stored metrics sent to the collector at once in bulk, or are they sent over time?
2) When the agent is monitoring multiple application pools and/or multiple executables, does each monitored app pool or executable have its own 15 second interval when sending metrics to the collector, or do they all share the same 15 second interval?
Any supported APM release
The agent metrics are placed in a queue in memory to be reported over to EM/collector. The queue would be cleared if the agent connection to EM was dropped and couldn’t be re-established shortly after retries. When connection is restored, agent will resume checking queue every few seconds to see if there are new items to be reported.
Application pools are independent, though they report at the same time. But its possible to see one start at 2:01 pm and another at 2:02 pm depending on when they are exercised and start to report. i.e. if you touch application A and a few seconds later you touch application B.