使用haufe.requestmonitoring条目分析zope实例日志
zope_lrr_analyzer的Python项目详细描述
目录
Introduction
此项目将向系统添加一个新的实用程序命令：zope_lrr_analyzer。此实用程序仅与 安装了haufe.requestmonitoring的zope实例日志（以及 monitoring long running requests hook已启用）。
因此，您的instance.log必须由如下条目填写：
------ 2012-03-27T15:58:19 WARNING RequestMonitor.DumpTrace Long running request Request 28060 "/VirtualHostBase/http/www.mysite.com:80/mysiteid/VirtualHostRoot/myrequest/..." running in thread 1133545792 since 10.7206499577s Python call stack (innermost first) ... lot of lines, depends on Python traceback ... Module ZPublisherEventsBackport.patch, line 80, in publish Module ZPublisher.Publish, line 202, in publish_module_standard Module ZPublisher.Publish, line 401, in publish_module Module ZServer.PubCore.ZServerPublisher, line 25, in __init__ <BLANKLINE>
该实用程序将帮助您分析收集一些统计数据的长时间运行的请求。
How to use
Usage: zope_lrr_analyzer [options] logfile [logfile…]
Analyze Zope instance log with haufe.requestmonitoring entries
- Options:
--version show program’s version number and exit -h, --help show this help message and exit -s START_FROM, --start=START_FROM start analysis after a given date/time (format like “YYYY-MM-DD HH:MM:SS”) -e END_AT, --end=END_AT stop analysis at given date/time (format like “YYYY- MM-DD HH:MM:SS”) -l LOG_SIZE, --log-size=LOG_SIZE keep only an amount of slow requests. Default is: no limit. -i INCLUDE_REGEX, --include=INCLUDE_REGEX a regexp expression that a calling path must match or will be discarded. Can be called multiple times, expanding the set -t TRACEBACK_INCLUDE_REGEX, --traceback-include=TRACEBACK_INCLUDE_REGEX a regexp expression that the Python traceback must match or will be discarded. Can be called multiple times, expanding the set -r, --keep-request-id Use request and thread ids to handle every match as a different entry
Results
让我们解释一下结果：
Stats from 2012-11-14 00:02:07 to 2012-11-15 09:55:41 (347 LRR catched)
...
----
2 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/foo/bar
25 - 3654.05561542 (1:00:54.055615) - from 2012-11-15 07:48:10 to 2012-11-15 08:45:29
----
1 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/baz
77 - 16029.3731236 (4:27:09.373124) - from 2012-11-15 07:43:55 to 2012-11-15 08:45:30
您将得到最慢的请求路径（前一个是最快的，后一个是最慢的）。 顺序是通过收集所有执行到同一路径的请求，然后获取总时间来完成的。
这意味着只调用一次需要30秒的请求比另一个路径快 这只需要10秒，但调用10次（30x1<；10x10）。
如果同时使用--keep-request-id选项，则每个请求都将作为单独的条目计算， 所以输出会有一点变化：
Stats from 2012-04-27 00:02:07 to 2012-04-27 16:55:41 (347 LRR catched)
...
----
2 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/foo/bar
1510.2860291 (0:25:10.286029) - from 2012-09-19 08:36:27 to 2012-09-19 09:01:22
----
1 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/baz
1750.49365091 (0:29:10.493651) - from 2012-09-19 08:30:34 to 2012-09-19 09:00:58
Single entry meaning
每个条目都提供这种数据：
Entry position Called path
| |
1 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/...
15 - 171.913325071 (0:02:51.913325) - from 2012-09-19 08:30:34 to 2012-09-19 09:00:58
| | | | |
Times called | Time needed (human readable) | |
| | Slow request end date
Time needed (in seconds) Slow request start date
使用--keep-request-id时：
Entry position Called path
| |
1 /VirtualHostBase/http/yoursite.com:80/siteid/VirtualHostRoot/...
1750.49365091 (0:29:10.493651) - from 2012-09-19 08:30:34 to 2012-09-19 09:00:58
| | | |
Time needed (in seconds) | Slow request start date |
| |
Time needed (human readable) Slow request end date
请注意，“time needed”信息是计算机计算时间。
Changelog
0.5 (2016-06-13)
- 显示捕获的lrr总数的信息 [基尔]
- 现在可以使用-s和-e作为日期字符串（不需要时间） [基尔]
0.4 (2012-12-06)
- 添加了--traceback-include选项 [基尔]
0.3 (2012-11-15)
- 始终显示请求的开始和结束日期 （不仅当-r选项被给定时） [基尔]
0.2 (2012-09-19)
- 添加了--keep-request-id选项 [基尔]
- 还存储（并在给定-r选项时显示）开始和结束请求时间 [基尔]
0.1 (2012-04-27)
- 首次发布
欢迎加入QQ群-->： 979659372
