High concurrency in ngnix cluster

server {

    listen 80;
    server_name manage.jt.com;
    location / {
    # agent 
    #proxy_pass http://localhost:8091;     
    proxy_pass http://jtWindows;
    }
}
#
upstream jtWindows {
# polling , Heavy ,iphash  down attribute   Hang up means to hang up   #backup  Spare machine 
    #ip_hash;
    #server 127.0.0.1:8081 weight=6;
    #server 127.0.0.1:8083 weight=3;
    #server 127.0.0.1:8082 weight=1;
    server 127.0.0.1:8081  down;#down attribute   Hang up means to hang up 
    server 127.0.0.1:8083  max_fails=1 fail_timeout=60s;# explain :max_fails  Indicates the number of failures , According to the regulations fail_timeout Of the time frame of          # Inside , If         # If the detection server fails, it will be in 60 No requests will be sent to this station in seconds tomcat.
    #60 It's going to be detected again in seconds .
    server 127.0.0.1:8082 backup;#backup  Spare machine