Summary: I need to log usage of resources on Windows XP machine to find if the lack of resources may be the reason for the application crash. I need your advice how to do that.
Context: There is an industrial PC with Windows XP, MS-SQL Server 2008 R2 Express, and the third party application.
The application collects data from the dedicated hardware, and stores the sensor values into the database. Basically, the data is the UTC and four temperatures. The sampling interval is 1 minute (i.e. kind of minimal load). To emphasize again, the application saves 5 values to the database table in one record only once a minute.
My task was to implement the SQL Server Service Broker solution to transfer the collected data to the central SQL server. (A trigger for the new record, XML formed, sent to the other side.) It works smoothly when it works. However...
The problem: It happens occasionally that the application stops to collect data. They claim that this is because there is not enough resources after the SSB solution was implemented. My guess it that they use the change at SQL server ony as a stock excuse. Anyway, they may be right. I need some exact proof what happened in the time of the crash.
As the application records UTC, I know when it stopped to work (from the missing temperature records).
My question: How can I log the usage of the resources, what resources should I observe, and what tool should I use for the purpose?
I am not trained in hardware/OS administration (just some general knowledge), and I have never solved that kind of problem before.
I know there is something named performance counters, but I did not use them, and I do not know how to log them. There is the powershell utility installed on the machine if it matters. (I have heard it is possible to access performance counters from the powershell, but I know nothing more.) I know there is perfmon on the machine, but I do not know if it can be used for the purpose.
Thanks for any suggestion on how to solve the problem,