Can anyone confirm what a load balancer should be using as a health check to determine if Volt MX Foundry is up and healthy on either server in the cluster?
Currently we are using Apache HTTP servers for reverse proxy and load balancing and we have this configured as (port numbers replaced with xxxx for security in this post):
ProxyRequests OffProxyPreserveHost on
SSLProxyEngine on
SSLProxyVerify none
SSLProxyCheckPeerCN off
SSLProxyCheckPeerName off
SSLProxyCheckPeerExpire off
ProxyPass /services balancer://integrationcluster/services stickysession=JSESSIONID
ProxyPassReverse /services balancer://integrationcluster/services stickysession=JSESSIONID
<Proxy balancer://integrationcluster>
BalancerMember https://10.137.4.68:xxxx route=server1
BalancerMember https://10.137.6.26:xxxx route=server2
But we're seeing some odd issues where traffic is still being sent to a server which ran out of heap memory and became unresponsive. We thought this traffic would go to the other healthy server.