Brian Robbins

@ivberg, this is pretty great!

Yeah, this looks like it's due to the fact that profiling is enabled. You can confirm that something else isn't causing this by explicitly disabling these events. With Perfview, try...

No, it should not be. I think that this environment uses a cached version of PerfView that used to do this, but the latest version should not.

@adamsitnik, sorry that's because the command I gave you shutoff unwind info publish. Try `/ClrEvents:JitSymbols`. Also, it looks like @sebastienros has updated to the latest version of PerfView, so it...

Since the PID is 1, I'm wondering if you're running this inside a container. That may have something to do with this. Can you tell me more about the environment?...

This text means that there were lost events during the trace, which does make the capture suspect: ``` Warning: Processed 2682678 events and lost 89 chunks! Check IO/CPU overload! ```...

Yes, it could. Can you try capturing with this private copy of perfcollect and see if that solves the problem? https://raw.githubusercontent.com/microsoft/perfview/04663ba0aa85c90e832f1e5a2ecb2fef972f2796/src/perfcollect/perfcollect

I suspect that these are two different issues. The hang is likely just the fact that there is a lot of code being used in the app. The more managed...

0x8 is `CooperativeBlocking`. This should be specific to .NET Core. The parser has been updated with this value at https://github.com/microsoft/perfview/blob/main/src/TraceEvent/Parsers/ClrTraceEventParser.cs#L12873.

CooperativeBlocking is a behavior that allows the threadpool to detect behavior that looks like sync-over-async and increase the rate of thread injection to avoid thread starvation.