Windows monitor not capturing some measures
I have configured the Windows monitor, it connects and retrieves some of the measurements but there are number that don't return. Looking at the details the measurements show the following error:
"Invalid data from Performance Counters. Please check if the host is reachable and the requested Performance Counter is available. Detailed error message: The returned data is valid"
I logged into the system we are attempting to monitor using the connection user we specified in the monitor setup, and ran perfmon. I was able to add a couple of the problematic measures and see data (I only tried a couple). I also double checked the counter names for differences including case, and they matched.
I'm stumped. Any ideas would be much appreciated.
See the screenshot below:
Pat,
you probably want to take a look at this article: https://apmcommunity.compuware.com/community/display/DOCDT41/How+To+Retrieve+Performance+Metrics+with+Windows+Performance+Monitor
Retrieving Windows Perfmon data can fail due to many reasons. Especially if it's done via remote authentication. The article mentions not tu specify a user in the actual monitor configuration but to make sure the executing collector runs under a user account that has privileges to remotely query a machine.
Also be aware of that Windows Perfcounter names are (sometimes) inconsistent and dependent on the OS version/patchlevel/security configuration/even language of the system and not very reliable.
Due to those reasons support for remote monitoring will be deprecated in the future and likely only be community supported. For system monitoring we will provide additional agents with the next release.
Reinhard
Thanks Reinhard,
I've actually already gone through most of that article... it's helped a lot in the past.
I should have mentioned that too: we ended up setting up the collector to run as a user with full permissions (including performance monitor permissions). It didn't help.
Does it do this every time? In other words, if you ran it again does it work? There are some counters that will fail the first time around because they are dependent on prior values (they are computations) and the first time around there are no previous values.
Thanks for the reply Rob. Yes I should have mentioned that, we manually ran it a few times with the same results.
