9
最近发现当upstream后端服务故障后，nginx的cpu占用就会飙升，后端服务故障的数量越多，cpu飙升就会越高。当后端有超过接近10台服务故障时，8个进程的nginx，有之前平均每个进程5-10%的cpu占用，直接飙升到99-100%，当注释掉故障后端服务后，cpu占用马上就正常了。 配置大致如下：
upstream https_backend {
consistent_hash $real_uri;
server 172.29.184.4:40443 weight=100;
server 172.29.184.5:40443 weight=100;
...serverlist...
check interval=3000 rise=2 fall=5 timeout=1000 type=http;
check_keepalive_requests 100;
check_http_send "HEAD /xxxx.jpg HTTP/1.1\r\nConnection: keep-alive\r\n\r\n";
check_http_expect_alive http_2xx http_3xx http_4xx;
}