We run websites on-premise and want to check the performance and availability from internet. So we use the PRTG monitoring tool with a "remote probe" on an Azure VM.
However on most HTTPS URLs that are tested we have the following intermittent error (in the PRTG monitoring):
Error establishing secure connection [TLS10-12] Error connecting with SSL. EOF was observed that violates the protocol [TLS10] Error connecting with SSL. EOF was observed that violates the protocol [SSL3] Error connecting with SSL. EOF was observed that violates the protocol [SSL2] Error connecting with SSL. EOF was observed that violates the protocol
This error only occurs once about every 100 connection attempts, so it is nearly impossible to nail down the source. And it does appear only from an Azure VM (Windows 2012-R2, A2 and A3 size, standard or basic makes no difference). If I move the same monitoring to a Amazon AWS VM or a PC on a home internet connection this error does not occur in the monitoring.
The errors seem to indicate that some firewall or IPS device might be dropping the connection, however this only occurs on Azure. I can't imagine anything that could cause this very strange behaviour, yet it is happening. For now the monitoring is happening from a probe on AWS/amazon, but I really do look for an explanation.
The machine itself was created with the new azure portal, with a fixed/reserved public IP and only RDP traffic on a high port is allowed from 3 public IPs. All other traffic from the machine is outgoing (to the core PRTG for which it is the probe and to the websites to test for availability).
