I have a problem, I can’t completely solve it using Google technology.
I need to serve> 500 concurrent connections over SSL over SSL. Clients ideally connect to wss: //wss.mydomain.com and get an instance of the instance.
Option 1 . I am currently using a custom managed virtual machine to deploy a Docker file that launches an SSL enabled web server server. Deploying an instance of a managed virtual machine results in a single GCE instance. I use Google Cloud DNS to point wss.mydomain.com to this one GCE instance IP address.
Problem: I cannot connect to wss.mydomain.com (even my-mvm.appspot.com), because connections through a web connection (via SSL) are not redirected to an instance of a managed virtual machine. Raised the issue , but did not see the answer that helps me. Others have also come to this conclusion here and here .
So, if managed virtual machines are not the answer, where can I find it in Google Cloud Architecture?
Option 2 . Since managedVM looks like a gluing coating for GCE, maybe I can save the current setting, which already provides my GCE instances, and add the load myself?
Option 2a . HTTP (S) load balancing does not allow web memory traffic, so one out .
Option 2b . Network load balancing can work. I need to assign instances to the target pool for the job. Can newly created virtual machine instances be assigned to the target pool so that they connect to load balancing? If so, will this be a complete solution to the problem?
Option 3 . I want to continue to use my Docker file for configuration. So I look at the container engine. I have no experience here at all.
Before I start getting into another minefield, which option is most likely to provide a solution to my problem? Perhaps there is another option that I did not think about.