I'm setting up a load balancer with nginx, where the requests made to the nginx server are distributed between two tomcat servers. I am using the ip_hash option to be able to work with login in the applications. However, when I close the application on one of the servers, the requests made to this server takes +/- 30 seconds. With each click it is necessary to wait this long period of time.
I would like to know if it is possible to have all requests redirected to the other server where the application is 100% functional after a short time waiting for server response.