IIS 8.5-hosted application on a dual-Xeon E5 2620 v2, 64 GB of RAM, and a 1 TB hard drive.
Zzz-11-Zzz Server Technology Client Side: 15 to Zzz Other Tools: After consulting some posts, we decided to remove Automapper from the continuous calling technique for Autofac 3.5, Automapper, and Newtonsoft JSON.
When 100 user are online, the CPU will use almost 100%. I tried a lot of debugger programs and performance monitors in an effort to find a solution, but to no avail. I did, however, generate some reports that appeared to be the root of the issue. Please see the image below for the result I received using certain debugging tools.
but have yet to pinpoint the issue,
We have a hosted app and see this issue; we believe it to be a coding and optimization issue. We have therefore begun to optimize the program and regularly upgrade the server version while monitoring usage, but there has been no improvement in performance.
It will get two requests every second and one input operation that is not based on a continue basis for each user based on their request. It can take a minute or longer.
When there are 100 users, one request takes roughly 500 to 800 milliseconds to complete after 100 user start the operation above, and as the number of users increases, the app becomes slower. Once there are 200 users, it will finish in 2 to 4 seconds. and get slower and slower.
According to a friend who also has the same application and user request, they have published on a VPS and are able to support 2000 users concurrently without experiencing any downtime. In contrast to them, I have a dedicated server with no resource sharing. I therefore believe that I will perform better than my friend.
According to some experienced friends' opinion, a request is being blocked, either in EF or iis. nonetheless, how and why. Every code is operating as usual.
Since it is currently causing me headaches, I am hoping to receive a remedy as soon as possible.
I appreciate it.
Your server is continuously being monitored by a plethora of technologies that affect throughput. Remove them from the production environment if you can. Although installing such tools in production situations is not advised, you can do so in test environments.
Please also describe your testing process for 200 or 100 queries. What programs are you employing to send 100–200 requests?