tag:blogger.com,1999:blog-17672034.post7042647822624692200..comments2018-05-25T03:15:14.811+12:00Comments on Bohdan Szymanik: Response Time Distributions, IIS Log Files, and the question of the Missing EventsBohdan Szymanikhttp://www.blogger.com/profile/13123728021973733938noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-17672034.post-42387897659058684322007-03-18T10:40:00.000+12:002007-03-18T10:40:00.000+12:00Thanks Max, we've done the same - we're now using ...Thanks Max, we've done the same - we're now using the times we obtain through an async call out to an event auditing database. We record the arrival time and processing completed time.Bohdan Szymanikhttps://www.blogger.com/profile/13123728021973733938noreply@blogger.comtag:blogger.com,1999:blog-17672034.post-52423678237506937552007-03-17T21:00:00.000+13:002007-03-17T21:00:00.000+13:00Same problem. Include any file upload and you get ...Same problem. Include any file upload and you get an even more skewed picture. I stopped relying on IIS logging for app profiling. What I do is an in-memory log that is flushed into a file on a schedule. The whole thing runs on a cluster and any given session can be spread accross any number of machines. An attempt to reconcile the app log with ISS by time proved to be inaccurate and I had to rely on other request parameters like the query string and user IP. <BR/><BR/>Then I just stopped using IIS logs for profiling and use the app logs instead. It's a bit of an overhead on the resources, but it's worth it. The only trouble is it's hard to track other resources that are served by IIS directly.<BR/><BR/>Cheers,<BR/>Maxweb2print guyhttps://www.blogger.com/profile/05905971097484296565noreply@blogger.com