Powerbi analyzing data for Windows performance counters

Tag: Cloud Platform for Cloud computing

When we do a system performance analysis or SQL profiling, we use one of the most primitive and useful tools, which is "Performance Monitor", so today I'll show you how to use Powerbi to analyze performance counters in Performance Monitor to analyze the performance of the system. Of course, after learning this you can extrapolate, to staging other performance, such as SQL performance, etc.

First, I'll take the Windows Performance Monitor sample to collect some of the performance counter's data

After sampling, I can get such a performance counter file

Next I'll use the Relog.exe program to convert this performance counter file into a CSV file

Then I use "notepad++" to open the converted file Systemperformance.csv, you can see each data value is separated by commas

The next step is to open the Powerbi and import the SYSTEMPERFORMANCE.TSV formatted data.

Select Systemperformance.csv After

Select Edit

Next, I'll remove the first line.

Input 1

Because each column is a name and multiple values in addition to the first column, then select the first column and right-click the reverse Pivot column

This will allow you to see the name and value of each one.

Next, in order to better identify the name of each property, next select the attribute column, select split Columns-by delimiter

Split by the \ delimiter in the attribute to determine

Next, remove the 2 columns of empty columns

Next, rename each column title

I'll split the object again and split the memory in parentheses into a column as an instance.

Input

Next, remove the new column, and select the new column for the replacement value.

Input) to determine

Next, we'll take this list as an example.

Next, you can close and apply

Wait for loading to complete

Next I create a line chart

Change dates from a date hierarchy to a date

Next, add a slicer, drag the object to the field

In the same way, add 2 slicers to the instance and type as fields respectively.

Click the instance slicer to format if you want the instance to have multiple selections instead of the default radio

Similarly, the type is also set to be multi-select and select all, so you can choose the object and type you want to view the data.

Next i import a visual view time Brush Timebrush, the time brush is used to choose to zoom in and out of the time interval of the visual view model, similar to a bit like audio editing software, to a section of the processing and amplification is a reason

Add this new visual view and set the following

Next, add an instance of the line chart to the legend of the line chart

Then I just select a time interval in the time brush, you can see in the line chart below this interval period of time to enlarge the value of the trend

Next, I'll copy an existing line chart, replacing the legend in this line chart with the instance type.

Finally, add 2 pie charts, each of which are instances and types of pie charts, which can be used to analyze the proportion

OK, I'll finish the layout again.

Because my data collection soon and is my laptop, so the ups and downs, see no obvious performance differences, but you can extrapolate on their own production server to analyze the performance of the system, of course, can also analyze other, as long as the performance Monitor can be captured can be analyzed, such as SQL , Exchange, and so on.

Powerbi analyzing data for Windows performance counters