UserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/68.0.3440.106 Safari/537.36

Example URL:

Steps to reproduce the problem:
I have a TCP loadbalancer (single IP) routing between two backend based on SNI (without SSL termination). Both backends terminate SSL and have same wildcard certificate while subdomains are different (e.g. one.example.com and two.example.com, with wildcard SSL certificate for *.example.com).

What is the expected behavior?
Requests for different hostnames should not reuse each others SSL sessions.

What went wrong?
Sometimes requests to one.example.com are routed to backend two.example.com and vice versa.
It happens when we try to reach two.* and one.* simoultaneously or in a short period of time (and from different tabs).
We can see in the loadbalancer log that the wrong request was initiated within previous SSL session, so SNI based routing was unable to detect the right destination.

Did this work before? N/A 

Chrome version: 68.0.3440.106  Channel: stable
OS Version: 6.1 (Windows 7, Windows Server 2008 R2)
Flash Version: