What documentation page is affected

https://www.elastic.co/guide/en/fleet/current/add-a-fleet-server.html

What change would you like to see?

We should include documentation about how any reverse proxy should be configured in order to support Fleet Server's long polling pattern with Agent. This typically requires configuring request read and response timeouts on any proxy or load balancer in front of Fleet Server.

• Describe the requirement in general
  • Agent makes long polling requests to Fleet Server to wait for configuration updates
  • These will take up to 5m on <= 8.6, and will default to ~25m in 8.7+
  • Things will still work when this is not configured correctly on the proxy, but performance may be degraded and load on Elasticsearch may be higher
• Include configs for examples for the most common proxies our customers use, such as Nginx, Apache, and K8s Ingress controllers.
  • Here's a theoretical example of how this would work for Nginx and Nginx Ingress Controller (still needs to be tested): https://github.com/elastic/sdh-beats/issues/2778#issuecomment-1340901379
• Include examples of common signals the user would find in the logs to identify that there may be a timeout issue occurring
• This may also affect proxies that Agents are using to connect to Fleet Server directly, and not just reverse proxies. We may want to include examples of how to configure these types of proxies too, such as Squid.

Additional info

No response