We are currently experience site outages on our site where latency spikes up and eventually outages occur. The site utilizes ASP.NET MVC/WebAPI with a SQL Server back end. It is hosted on AWS. All core resources on the server look fine (RAM, CPU, etc.) The one thing we have noticed is if you look at the "WorkerProcesses" option in IIS you see all the requests (100's) are hanging for between 20 seconds to over 2 minutes in a "SendResponse" stage. Doing research on the internet it seems like this is the point when IIS is attempting to send the response back to the client. Some have suggested that it is mobile users bringing down the server, but I am not sure that could be the case for us. For our setup their is a ELB load balancer between the client and the web servers. I was hoping someone could help point us in the right direction as far as either additional resources we could use to troubleshoot the problem, past experiences and things that would be helpful to diagnose the issue, or else suggested IIS tweaks that might help alleviate the issue.
Thanks, Kevin