How To Troubleshoot Intermittent Timeouts

Question

We have an intranet running Sharepoint 2007 on a single Windows Server 2003/IIS6. On one particular list, some of our users see System.Web.HttpException 'request timed out' YSOD pages, but the errors don't always occur in the same places. Some users see them when they are using the person lookup tool to select a person, some users see them when they try and save a new entry to the list. Some users get errors some days but not others. Occasionally a user can save an entry to the list without seeing an error page, but the entry doesn't then appear in the list.

Can anyone offer pointers to how I can troubleshoot this problem?

Answer 1

You need to understand what might be causing the timeouts (obviously!) - the problem is this could be almost anything. I would start by looking at perfmon counters, check to see if you generally have large request times and how many requests are queued.

A MS tool called managed stack explorer can be used at the times the issue happens to dump a list of what the active threads in the app pool are doing, you can often see blocking issues by inspecting that log too - be careful with this though, don't leave it running for long on a live system as it can be quite invasive.

You could for instance have long running calls which are causing thread blocking in IIS, this can often happen when server side code makes external calls to web services as an example.

Get someone with good SQL knowledge to set up alerts when sql blocking occurs, if it does then you might be able to track down specific areas of the site which cause blocking which in turns causes requests to queue up and then timeouts to occur.