cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Which metric in Windows Perf.mon represents "Hard page faults" in Appmon?

ali_fetouni
Contributor

Hello!

I have a customer that is asking which metric in Windows Perf.mon represents "Hard page faults" in Appmon? Is there any documentation around this?


Many thanks!

-Ali

3 REPLIES 3

shane_korbel
Dynatrace Pro
Dynatrace Pro

Hard page faults occur when a process refers to a page in virtual memory that is not in its working set or elsewhere in physical memory, and must be retrieved from disk. Dynatrace counts the number of pages read from disk to resolve page faults. This includes also pages read to satisfy faults in the file system cache (usually requested by applications) and in non-cached mapped memory files.

Window's Resource Monitor lists so-called "hard faults" per process. It seems that it counts page reads to satisfy faults in non-cached mapped memory files but it completely ignores pages read to satisfy faults in the file system cache. Since file system cache is global, a distinct assignment of page faults to single processes is not feasible. But instead of counting the faults for the virtual "System" process, Resource Monitor seems to ignore them completely.

I suggest to search for a process which either reads a very huge file (size > a standby memory) or reads a very high number of files within a short time period (e.g. to index all files on a local disk or a network share). The Disk Activity view in Window's Resource Manager may help here. Another helpful too could be RAMMap's File Summary tap (https://technet.microsoft.com/en-us/sysinternals/rammap.aspx)

ali_fetouni
Contributor

@Shane K., thank you for answering! I take if from your answer that there isn't a 1 to 1 mapping with regards to the mapping of the metric HPF (in AppMon) and in perf.mon (Windows)? And that the metric HPF itself is comprised of several elements that are represented as different metrics in perf.mon?

shane_korbel
Dynatrace Pro
Dynatrace Pro

The measure that you're looking at in Dynatrace is:

PageReadCount 

The number of pages read from disk to resolve page faults.

The PageReadCount is the equivelant of the Windows performance monitor counter "Pages Input/sec" - the amount of pages read to resolve hard page faults.