Two small details about the Web log statistical analysis

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

Webmaster friends will usually give their own website installation CNZZ, Baidu statistics, such as webmaster statistics tools, but, these webmaster statistics tools will not record the crawling of web spiders. Some webmaster friends usually use the log Analysis tool to analyze the spider crawl situation on the website log. I personally think that most of the webmaster to the site log statistical analysis process, ignoring a number of small details, the following list of two:

First, the site log file should be based on the number of visits to determine whether the need to build by the hour.

One of my webmaster friends, the site is generated every day a log file, some time ago he participated in the Electricity Business Circle competition, the site ranking to keep in the home page, daily traffic has thousands of IP, the daily site log file size about 50M, a bit of a cup with his computer is old Point, an open Web site log file is not no response or panic. He had to pass the log to me through the network, let me help analysis, 50M file Although not big, the problem is that he uses the telecommunication network, I use the network of Netcom, in the time of passing the log often cups. 50M log files, I used the log analysis program also often data overflow, crash, but have to use text editing to open the view, facing the dense log text, statistical analysis of these data is very very difficult. Therefore, it is recommended that the site to visit a larger webmaster friends, it is best to generate Web site logs by the hour, although the production of more than a point, but more conducive to the analysis of the site log.

Second, the site log records of the information is actually not complete.

Do not know webmaster friends, did not notice that the site log rarely out of 5xx return code. For example, the 500 return code indicates an internal server error, and a 503 return code indicates that the service is not available. Webmaster friends know that the 5xx return code generally means that the Web server has failed, under normal circumstances, the server failed to generate Web site log. In other words, when the Web server down, or DNS resolution, all can not access, spiders can not visit, in this period of time, the site log is certainly unable to record any information. In order to better monitor the site situation, I personally recommend that you register and use Google Webmaster management tools, can effectively record the server access error messages.

Above two points, is I personally to the Web site log analysis process to think of two small problems, hope to be able to make a point, welcome to Webmaster friends to shoot bricks.

This article originates from Beijing electric appliance Repair Net http://bbs.bjjdwx.com/thread-135013-1-1.html, reprint please indicate the source, thank the cooperation.