Server availability automatically becomes "unknown" randomly when they are shown as "Ready for load balancing" using WFF2.1

Setup is using WFF 2.1 with 1 controller (10.0.2.10), 1 primary (10.0.2.11) and 1 secondary (10.0.2.12).

Every so often, a few times a day, the website will just hit a 502 error and upon checking the Monitoring and Management console, the servers' availability will show up as unknown.

502 - Web server received an invalid response while acting as a gateway or proxy server.

The secondary server sometimes gets marked as unhealthy due to failure to run some sync tasks but there are no errors for the primary server logged in the WebFarmService.log so I'm curious as to why both servers are not able to serve any requests. This defeats the reason for using the WFF in the first place for the reason of better availability. Any help?

Both primary and secondary servers' availability shows up as unknown randomly for reasons unknown. (i.e. after returning to work on a new day to find this when no changes were made and both were confirmed to be available and serving web requests the evening before.)

http://i.stack.img开发者_JS百科ur.com/gjMsy.png

Both primary and secondary servers are ready for load balancing

http://i.stack.imgur.com/qESEN.png

---

It means that application pool is stopped (application wasn't used for more than 20 minutes). Then it becomes unknown.

502 is because of timeout, when you increase timeout on proxy server (controller, where ARR is) to 240 sec. it should work ok. Sometimes default 120 sec is too short to warm up app pool.