Am using Caddy
Have 1 unix Reverse Proxy, with 2 Aware servers, and another MSSQL data server backend (4 servers total)
I desire to have a health check which redirects traffic from Aware1 (primary) to Aware2 (backup) in the event of the primary going down.
The goal is to minimize user downtime in the event something happens to the primary Aware server.
But when traffic IS redirected AND A USER IS ACTIVE, its going to cause a login screen to appear (I would imagine - this isn't set up yet) due to the fact that the 2nd Tomcat server doesn't have a session for that Primary user.
So in the middle of work, the user will be asked to login again with no explanation.
And then once the Primary server is restored (lets say 5 mins, because an admin will be notified and can correct whatever happened) and the Health status is restored, THEN ONCE AGAIN the user(s) will have to get logged in as they all are automatically re-routed to the Primary.
Does this sound about right?
Does anyone not like this behavior? (I can imagine users will be annoyed if this happens too frequently, but its not like I have a server going down even once a week, so I'm not sure if it will even be a big deal)
Any way to improve on this scenario?
