OMS today released a new feature Performance monitoring. Not sure if that is the official name but basically allows you to gather performance data from servers by adding performance counters.
The feature is located under Logs as it represents different king of log gathering. Keep in mind that this performance data gathering feature is different than the Capacity solution. This feature does not require SCOM-VMM integration and works with agent-only connected computers.
When you navigate to Logs initially you will see a box which will help you to add the most common performance counters initially:
Of course you have the option to add all of them select only a few of them or add other counters.
When you click add you will see that besides that you can choose which performance counters to add you can also choose at what interval they are gathered. 10 Seconds is the minimum. I do not know that is the maximum but I’ve tried with 10 minutes and it works. How frequent performance data is gathered reflects in the accuracy of the data. If you gather data on 2 minutes for example you will not be able to see small spikes if you have them on your servers.
If you want to add other counters just type some word and a matched of performance counters for that word will appear. Very easy and convenient.
If you follow the format you can even add your own custom counters if you have them on your server/s. Just follow the format of the other counters if you are confused how to add them:
In summary the configurations are very flexible and convenient. Some performance counters may not be updated with new data every second thus you can gather data on longer interval and save money. I had different initial expectations on the configuration of this feature and now that is GA it exceeds them a lot.
Once you’ve enabled a few counters you will go to Search and type: “Type:Perf”
Below the search you will see View for Metrics. Click on it:
This brings cool graphics of your performance data that has been gathered. When you start initially you can narrow down the results for the last 6 hours:
This view gives you nice overview if you need to see overall your performance data as it allows you to see more graphics at one page. Each graphics can be expanded which will show you more details about that performance counter:
And actually you can expand all the graphics on the page if needed.
Depending on your interval of your data gathering you soon will see if you stage on the same page without refreshing it that new data is being visualized live:
With lighter blue color we can see live the new data coming.
Along with the graphics you will see Last and Average value being displayed as well on the right.
Now this performance data is available in OMS as logs that are aggregated and you can actually use with queries.
Type:Perf (CounterName=”% Processor Time”) | measure Avg(Average) as AVGCPUTime by Computer | Where AVGCPUTime>10 | Sort AVGCPUTime desc | Top 5
When you have more servers you will get better results with that query.
As this feature is part of Log Management you will see Log Management using more data in Usage:
But the feature is very light on the cost from my perspective and you can always increase the interval of data gathering. In my case Log Management is big because I’ve enabled a lot of logs.
I hope you will like this feature. I am pretty excited about it and it exceeds way more my initial expectations. Of course I’ve got some feedback as well which I’ve already gave to the product group. If you have any feedback as well feel free to share it on the UserVoice.