本文发表在 rolia.net 枫下论坛请各位高手帮我分析分析:
最近刚加入一家公司,第一个星期就碰到这个棘手问题, 我们SUPPORT的APPLICATION开始出现PERFORMANCE问题. Production environment - load balancer with two web servers, talk to application server, applcation server talk to SQL DB server. NET 2.0 web application, SQL Server 2000 backend database, around 4000 end users, most web pages are loaded with big chunk of data from complex query/stored procedures; application uses connection pooling (max pool size=1000,connect timeout=60) .
The problem occurs on 14th of every month, because emails with report links are sent out, many users click those links and login to the site and load reports, start from 10AM to 11AM 形成 peak hours, application becomes very slow or even not responding at all, some users get timeout errors, some get connection related errors.
从Web statistics上看(以下数据只是大概),
8 AM - 20 Hits Response Time: 3 secs
9 AM - 40 Hits Response Time: 6 secs
11AM - 156 Hits Response Time: 30 secs
12AM - 180 Hits Response Time: 45 secs
1PM - 110 Hits Response Time: 60 secs
2PM - 50Hits Response Time: 60 secs
3PM - 30Hits Response Time: 60 secs
4PM - 10Hits Response Time: 60secs
后来他们做了IISRESET,Server才恢复正常. 症状就是: web server 在peak hours 时很慢很慢,基本不工作, 奇怪的是 peak hours 之后Web Serer 还没有恢复正常.
请大家建议我应该怎么诊断好, 因为是production, 我能拿到的信息很有限. 如果要解决的话, 有没有好的solution?
谢谢!!更多精彩文章及讨论,请光临枫下论坛 rolia.net
最近刚加入一家公司,第一个星期就碰到这个棘手问题, 我们SUPPORT的APPLICATION开始出现PERFORMANCE问题. Production environment - load balancer with two web servers, talk to application server, applcation server talk to SQL DB server. NET 2.0 web application, SQL Server 2000 backend database, around 4000 end users, most web pages are loaded with big chunk of data from complex query/stored procedures; application uses connection pooling (max pool size=1000,connect timeout=60) .
The problem occurs on 14th of every month, because emails with report links are sent out, many users click those links and login to the site and load reports, start from 10AM to 11AM 形成 peak hours, application becomes very slow or even not responding at all, some users get timeout errors, some get connection related errors.
从Web statistics上看(以下数据只是大概),
8 AM - 20 Hits Response Time: 3 secs
9 AM - 40 Hits Response Time: 6 secs
11AM - 156 Hits Response Time: 30 secs
12AM - 180 Hits Response Time: 45 secs
1PM - 110 Hits Response Time: 60 secs
2PM - 50Hits Response Time: 60 secs
3PM - 30Hits Response Time: 60 secs
4PM - 10Hits Response Time: 60secs
后来他们做了IISRESET,Server才恢复正常. 症状就是: web server 在peak hours 时很慢很慢,基本不工作, 奇怪的是 peak hours 之后Web Serer 还没有恢复正常.
请大家建议我应该怎么诊断好, 因为是production, 我能拿到的信息很有限. 如果要解决的话, 有没有好的solution?
谢谢!!更多精彩文章及讨论,请光临枫下论坛 rolia.net