Recently, when building the site api, I found that if there are too many connections, the backend will fail to respond and crash directly, so the request should be limited. After some searching, I found the HttpLimitZoneModule module of Nginx.

# Use HttpLimitZoneModule to limit the number of concurrent connections

limit_zone can only be defined in the http scope, limit_conn can be defined in the http server location scope.

http {
   limit_conn_zone $binary_remote_addr zone = conn:10m;
   #Define a limit_conn_zone named conn to store sessions, the size is 10M memory, 1M can store 16000 states;
   #$binary_remote_addr is to limit the same client ip address;
   #$server_name is to limit the maximum concurrent number of the same server;

   limit_req_zone $binary_remote_addr zone=allips:10m rate=20r/s;
   #Define a limit_req_zone named allips to store sessions, the size is 10M memory, 1M can store 16,000 states, use $binary_remote_addr as the key, limit the average request per second to 20, the value of rate must be an integer, if the limit One request in two seconds, can be set to 30r/m

   server {

   location {
  limit_conn conn 20;
  #limit_conn is to limit the number of concurrent connections;

  limit_rate 500k;
  #limit_rate is to limit the download speed;

  limit_req zone=allips burst=5 nodelay;
  #Limit no more than 20 requests per ip per second, the number of leaky buckets burst is 5, and brust means that if there are 19 requests in the 1st, 2, 3, and 4th seconds, and 25 requests in the 5th second are blocked Allowed. But if you make 25 requests in the first second, and more than 20 requests in the second second return a 503 error. nodelay, if this option is not set, the average rate is strictly used to limit the number of requests. When there are 25 requests in the first second, 5 requests will be executed in the second second. If nodelay is set, 25 requests will be executed in the first second.

   }
}