We run a web application on a VPS (WinServer2012R2) DF18.2 The app allows patients to book appointments online.

We are regularly getting reports of patients turning up for appointments that do not show on the local medical practice systems. Examination of the "cloud" SQL database shows that the bookings have been successfully made. The problem has been traced to the fact that some of these bookings are not being transferred down from the webapp site to the local system.

Every 30 seconds, a web service call is made to check for any recent bookings and if found, transfer the booking data back down to the site. Our recently enhanced logs are showing that there are numerous "Http request failed" errors being logged on the local systems when this call is made.

There is no pattern. Sometimes the failures are successive (every 30 seconds), but there are also random minutes or even hours between failures. The bulk of calls succeed.

If EVERY call failed, we would be in a much better position to debug, but as is always the case, random failures have us scratching our heads.