I’m just wondering, did something change on how you are handling multipart webhook responses recently (last few months) in the particle API / on the particle cloud? Our fleet has lots of devices on 0.5.1 system firmware, and some on 0.6.3, which we can’t currently upgrade to newer system firmware due to memory constraints. Up until the last 2 months or so, we have had no issue with multipart webhook responses, but just recently there is data corruption a good proportion of the time.
From the details you have given it sounds like you are receiving responses out-of-order.
Particle recently added redundancy to Webhooks and this can cause responses to arrive out-of-order. For an explanation please see Webhooks | Reference Documentation | Particle
Thanks! This does sound like it is probably the cause of this issue.