Internal Error Server Point to Working Server Using Load Balancer Or Nginx

You avoid a white page by defining a properly formatted error document with the error_page directive instead of relying on the default error message :) On the other hand if you want to avoid sending further requests to a node that returns errors? Typically that is something you configure in the loadbalancer, there you want to remove a failed node from the pool of active back-end servers. The details differ between products but you need to define a health check that does more than just a TCP connect to determine if a back-end server is up and running. The loadbalancer needs to do something a little more advanced such as an actual HTTP request and evaluate how fast and what gets returned, typically if the HTTP status/response code is 200 and/or the response body has specific content or meets a certain minimum size.

Lets say that I have 3 physical servers with different IPs and 1 load balancer which manages the traffic between those 3 servers. We tested to stop the service of nginx in the 2 servers and the load balancer detected that and didn't use the 2 servers with nginx stopped.My question is:For instance that the 3 servers are up and running, and 1 server has an internal server error (any 500 internal server error). How can we avoid that 1 server that is encountering the server error? to avoid white page?

·OTHER ANSWER:

Lets say that I have 3 physical servers with different IPs and 1 load balancer which manages the traffic between those 3 servers. We tested to stop the service of nginx in the 2 servers and the load balancer detected that and didn't use the 2 servers with nginx stopped.My question is:For instance that the 3 servers are up and running, and 1 server has an internal server error (any 500 internal server error). How can we avoid that 1 server that is encountering the server error? to avoid white page?