如何调试Heroku请求超时错误 [英] How to debug a Heroku request timeout error

问题描述

我如何找出导致heroku上h12超时错误的原因？

How would I find out what is causing h12 timeout errors on heroku?

它是在不同页面/控制器上随机发生的。

It happens randomly on different pages/controllers.

这是我从日志中得到的错误

This is the error that I get from the logs

Processing by UsersController#new as HTML
2013-08-15T13:08:54.718157+00:00 app[web.1]: Started GET "/signup" for 105.226.239.198 at 2013-08-15 13:08:54 +0000
2013-08-15T13:08:54.744390+00:00 app[web.1]:   Rendered communal/_error_messages.html.erb (0.6ms)
2013-08-15T13:09:24.714225+00:00 heroku[router]: at=error code=H12 desc="Request timeout" method=GET path=/signup host=www.goodwordlabs.com fwd="105.226.239.198" dyno=web.1 connect=3ms service=30000ms status=503 bytes=0
2013-08-15T13:09:25.518168+00:00 app[web.1]: E, [2013-08-15T13:09:25.517428 #2] ERROR -- : worker=0 PID:5 timeout (31s > 30s), killing
2013-08-15T13:09:25.554402+00:00 app[web.1]: E, [2013-08-15T13:09:25.554166 #2] ERROR -- : reaped #<Process::Status: pid 5 SIGKILL (signal 9)> worker=0
2013-08-15T13:09:25.980010+00:00 app[web.1]: I, [2013-08-15T13:09:25.979551 #26]  INFO -- : worker=0 ready
2013-08-15T13:08:51.088169+00:00 heroku[router]: at=info method=GET path=/assets/clients/client1-03dfd2afecf092e15b9e3350d04acdc6.png host=www.goodwordlabs.com fwd="105.226.239.198" dyno=web.1 connect=2ms service=7ms status=304 bytes=0

推荐答案

以下是应该如何继续

Here is how you should proceed

1. 安装应用程序监控解决方案 newrelic 。它会给你的URL列表超时。新文物已付费，但也有新文物的免费版本。


2. 现在，尝试隔离&重现超时错误。因为如果你不能重现它。您无法修复它。

1. Install a application monitoring solution newrelic. It would give you list of URL that are timing out. New relic is paid, but there is free version of new relic too.
2. Now, try to isolate & reproduce timeout error. because if you can't reproduce it. you can't fix it.

这篇关于如何调试Heroku请求超时错误的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！