nginx+lua性能问题

ruochen

这是nginx的access.log，最后一列是request_time

[14/Aug/2013:18:04:36 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.739"
[14/Aug/2013:18:05:31 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "3.423"
[14/Aug/2013:18:06:21 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.751"
[14/Aug/2013:18:06:31 +0800] 123.147.244.64 - - "GET /newstatus?userid=smdnCZR8J7&code=cfouzZGEU0 HTTP/1.0"200 214 "-""-" "-" "5.010"
[14/Aug/2013:18:07:07 +0800] 117.136.8.80 - - "GET /newstatus?userid=ywgP7hekno&code=6pa85EWpDl HTTP/1.0"200 214 "-""-" "-" "5.006"
[14/Aug/2013:18:07:13 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.777"
[14/Aug/2013:18:08:07 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.579"
[14/Aug/2013:18:09:00 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.521"
[14/Aug/2013:18:09:53 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.670"
[14/Aug/2013:18:10:45 +0800] 117.136.8.56 - - "GET /newstatus?userid=vpH3qYJFMa&code=oQKcGPP1e3 HTTP/1.0"200 214 "-""-" "-" "0.554"

newstatus是lua写的，已经开启了lua cache，lua的处理逻辑我们打印了处理时间，都在10毫秒以内，初步怀疑是网络不稳定造成的
这些是2G终端在nginx的access.log请求记录
如何确认这样的性能问题？  tcpdump？有其他更简单的方法或者工具么？

数据流：终端---http（上报数据）---nginx-----lua---处理和保存数据---http（推送数据）---终端

laputa73

用替换法.
  终端每次都请求一个helloword的简单逻辑或者一个静态页面.
同时用一个同网段的终端进行测试