we are using ServiceStack + Redis + mySQL, it installed on Debian 8 + ngnix + fastCGI, as an installation guide we used the article: http://www.jokecamp.com/blog/servicestack-api-with-fastcgi-mono-server-and-nginx-hosted-on-digitalocean/
when we testing server. By increasing the load on the server at some point the server don’t answer on requests and stop. When we restart ngnix everthing begin to work correctly.
Could you help us, where can we get some information about server configuration to withstand heavy loads.
We’ve since published our recommended setup with Mono on Linux at: https://github.com/ServiceStackApps/mono-server-config - using HyperFastCGI should minimize some of the issues, but ultimately the load issues with Mono are a well-known limitation of Mono’s poorly maintained HTTP Stack and is not something which can be fixed within ServiceStack which is why we don’t recommend using HTTP on Mono for heavy workloads.
We expect the future for hosting .NET on Linux is going to be CoreCLR whenever that ships as it will be supported by Microsoft and have dedicated resources behind it improving any performance and stability issues. You can vote for it on UserVoice to get notified of any updates on it.
Tell me please, is it better to use Apache mod_mono then HyperFastcgi for heavy loads?
Heavy Loads should use Windows. But on Mono/Linux our preference is nginx/HyperFastCGI.