I had a look at all this new info but alas I don't think that's what is happening in our case.  I took some code that independently generates an HMAC (https://gist.github.com/andjosh/5c4f0244914adfd312e4) and for the same webhook, it's generating the...

Did you every get to the bottom of this?  I'm just looking into the same issue where a small number of webhooks come in and get a mismatch in the hmac (yet the code is the same that computes the hmac).  By small number, it's around 20 out of 1500 rec...