How to investigate upstream timed out issues?

[michael.heuberger]

Hello guys

In my nginx version 1.3.14 I'm having lots of upstream timeouts like these
and wonder what's the correct, professional approach is, to solve these:

example:

762#0: *113 upstream timed out (110: Connection timed out) while reading
response header from upstream, client: 58.28.152.233, server: videomail.io,
request: "GET /socket.io/socket.io.v0.9.11.js HTTP/1.1", upstream:
"https://127.0.0.1:4443/socket.io/socket.io.v0.9.11.js", host:
"videomail.io"

relevant nginx config:

  location / {
    proxy_cache one;
    try_files $uri @proxy;
  }

  location @proxy {
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header X-NginX-Proxy true;
    proxy_set_header Host $http_host;

    proxy_next_upstream error timeout invalid_header http_500 http_502
http_503 http_504 http_404;

    proxy_pass https://127.0.0.1:4443;

    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";
    proxy_http_version 1.1;
  }

what would you do in my case to examine this? imo the default timeout of 60
seconds is good and i do not want to increase it. it should be able to
deliver socket.io.v0.9.11.js under 60 seconds, so i believe the problem is
something else. i just do not know how to investigate this.

any hints very welcome!

cheers
michael

Posted at Nginx Forum: http://forum.nginx.org/read.php?2,237385,237385#msg-237385