nginx - apaceh concurrency issues

Question

Requirements: Apache, as a web server, receives REST requests. The peak concurrent request volume is 500qps, the average value is 50qps, and the time performance requirement is within 50ms.

The current problem:
The time from request initiation to program execution is about 20ms:
The average program processing time is: 6ms
The callback time is about 1ms

There is no problem in processing according to the average, but there are often some 100+ms requests, accounting for about 1%, please provide an optimized line of defense

Answer 1

This may involve your machine configuration, processing performance, bandwidth, data operations in the business, etc. If the concurrency is large, you can consider load balancing and try it