I have a REST service that I have exposed via a Site called public. It has a URL like this:
This service is use to receive callbacks from an external system and it has been working for over 2 years.
It has recently started missing come of those callbacks. Not all of them - just some of them.
I did the obvious thing, and added lots of
try..catch, logging, etc. but there was no trace of the missing callbacks.
I opened a Partner Support Case (still open) and they were able to provide logs of incoming calls to my endpoint. There were logs for the periods when it was working. An no logs in the time periods where we are missing callbacks - which seems to rule out bugs in my code, because they would still have a log even if my code has crashed.
The system calling my endpoint has provided logs which appear to show them receiving a 200 status code and
OK body. Which is what my Apex returns on success. Salesforce have no logs for the corresponding time on their side (we've checked timezones, and even used a window of a whole day).
It might be a coincidence, but this seems to have started at the time that our org migrated from na24 to na80. Really, this shouldn't make any difference as the callback URL doesn't include the instance - it uses the Site URL.
Has anyone seen anything like this? Got any ideas?