Nick Banks

@LPardue that works for me.

https://github.com/microsoft/msquic/blob/main/src/inc/quic_trace.h#L44 ```C typedef enum QUIC_FLOW_BLOCK_REASON { QUIC_FLOW_BLOCKED_SCHEDULING = 0x01, QUIC_FLOW_BLOCKED_PACING = 0x02, QUIC_FLOW_BLOCKED_AMPLIFICATION_PROT = 0x04, QUIC_FLOW_BLOCKED_CONGESTION_CONTROL = 0x08, QUIC_FLOW_BLOCKED_CONN_FLOW_CONTROL = 0x10, QUIC_FLOW_BLOCKED_STREAM_ID_FLOW_CONTROL= 0x20, QUIC_FLOW_BLOCKED_STREAM_FLOW_CONTROL = 0x40, QUIC_FLOW_BLOCKED_APP = 0x80 }...

And the connection has a set of flags, as well as each stream has its own set.

Personally, I think events **should** be tied to protocol, and "sharing" them should not be a goal. IMO that's going to make things harder and more confusing to interpret, especially...

And one more data point for you, we generally follow a pattern for events to be named something along the line of , such as `QuicConnVersionSet` and `QuicStreamCreated`. It allows...

Personally, I'd add a PathCreated event, that defines an arbitrary ID and all the parameters necessary to describe it. You might also want a PathUpdated event if you need to...

You obviously cannot just use the ID value. You must search for two paths with the same parameters and associate those path objects.

If you're warry of PII, you simply include a hash of the information. Then your parser must understand both raw data format and hash format. If one side includes raw...

First off, I think the notion of privacy should be a separate Issue and not discussed here. Still, I don't think you're going to be able to completely solve this...

In general, I agree a single new field seems reasonable, but which events does it apply to? Just packet events? Congestion control / loss recovery? Not every event needs to...