Additional Diagnostics

[davidfowl]

• We should add an event that emits the Kestrel settings. Specifically the configured KestrelServerOptions settings (Log KestrelServerOptions when tracing is enabled #34680)
• We should add an event that emits the IIS settings.
• We should add an event that emits the HTTP.sys settings.
• MVC options including what application parts were detected (@pranavkm)

Potential other diagnostics:

• Kestrel Memory Pool utilization vs. size
  • Counter for total blocks in bytes
  • Counter for checked-out blocks in bytes
  • Event for newly allocated blocks
  • Counter for newly created pool instance (could be event too)
  • These together can be used to help us decide how much we should invest in Allow server Memory Pool to shrink #27394 which is to "Allow server Memory Pool to shrink"
  • If we do allow the memory pool to shrink, we could add an event for shrinking. We could use these events to determine how many extra allocations the shrinking caused in a given app.

similar issue in runtime repo: dotnet/runtime#56072

[BrennanConroy]

@davidfowl what's the plan here?

[davidfowl]

To do the thing well.

[Pilchie]

@davidfowl - when is more coming here? It's getting quite late.

[davidfowl]

I think for RC2, I'll see if I do the low hanging fruit one. Maybe just HTTP.sys. The memory pool one needs more work.

[adityamandaleeka]

@davidfowl Any update on this one?