I am seeing the same thing as others here:

Windows, UDP packets silently dropped

And

Python dropping UDP packet that WireShark sees

(Windows 10, Python 3.6) - in my case I am talking to a device (Embedded system running LWiP) I send it a series of commands (about N=100), each command has a UDP response, commands tend to be short, mostly lessthen 20 bytes, responses are the same size, but some are huge (1200 bytes)

Occasionally Python does not receive a response UDP packet - yea, I know UDP is not reliable {flame: ON, please dont go ther with me, flame: OFF}

1. I'm on a private network in a lab at work, there is minimal traffic.
2. On the windows box (via wireshark) I see the response
3. BUT - python does not see the response.
4. The damn system worked this morning, and does not this afternoon.
5. This has happened before, and it goes away - and then it decides 'F-you' it is thursday, It does not work today, just because ...
6. I cannot change 5 man years of development at this point, lots of stuff would change, embedded code, fielded devices, test automation, validation scripts - tons and tons of Python - that would be a historic level of effort that nobody will accept.

I need the reason for this drop to be far more reliable or understandable, and not drop packets for no damn reason. I have no simple means to debug why my pythons script does this (and I assume it is python, I can't prove this.

on Linux - I could do that - on Linux I would use STRACE() and trace the system calls, and see the return values in the output log - but this is windows, there is no strace()

More about the issue:

I talk to a LWip STACK, In a loop I send it a query, give me data items X to Y, the device responds, then I ask for the next range - depending on what is being queried, there are 2 to 3 loops before it responds with an empty list.

Then it suddenly says: "Nope not going to deliver that packet" I need to understand why :-(