I'm trying to track down why our servers become unresponsive over time. After about 3-5 hours CF will start to queue requests and the only thing I can do is restart the services. I'm not seeing any hints in all of our hardware monitoring that it is a hardware related issue. Our database servers are fine, the load on the systems are fine but after about 3-5 hours, they'll just start queuing and are unable to recover on their own without a restart. The only hint I've been able to gather is there will be a large number of locked threads.
↧