Utilizing the Windows Performance Monitor for DIY Benchmarking and Logging
.jpeg "windows-performance-monitor-diy-benchmarking-logging")
- Learn how PerfMon can help you start making guesses about PC problems and then start knowing the real data.
- Get familiar with basic, rational performance counter language so that you know what you are monitoring.
- Learn the big distinction between observing a real-time chart and configuring automated monitoring to gain the real knowledge.
- PerfMon, even though you may never have used it, still, follow a clear walkthrough to write your first automated Data Collector Set.
- Learn how to read the graphs and report and make sense of the lines on a chart into understandable pieces about your system.
- Come up with methods of interpreting data recorded in logs, identifying patterns and relationships between symptoms and underlying causes.
- Train to monitor the things that are important to you specifically and not the generic system statistics.
- Get best practices to structure your logging in such a way that it is useful, not overwhelming.
- Learn how to trace various measurements, such as memory and disk activity, so that you can locate the actual cause of a slowdown.
- Establish a personal performance baseline a digital fingerprint of your PC during its health as your ultimate point of reference troubleshooting.
- Take reasonable recommendations on how to handle each of the log files and storage to maintain a light monitoring.
Raise your troubleshooting to more than a quick fix.
Have you sometimes been annoyed by a computer suddenly sluggishness, watched a spinning cursor, or been confused by a random crash in an application? You're not alone. Most of us have been there. We launch the Task Manager and are presented with an incomprehensible list of numbers, and we usually close the task manager more perplexed than we were before. This is a reactive attempt at attempting to resolve a problem whilst it occurs which is similar to attempting to comprehend a storm through a single raindrop. There must be a means of observing the trends with time.
It is this power that I have written this guide to achieve in you. We will be discussing one of the tools that you are already familiar with which is the windows performance monitor (PerfMon). This is not only limited to the IT professionals, PerfMon is an immensely useful companion to any computer user who wishes to know their machine more. You will know how to collect facts, as opposed to having a vague feeling concerning performance. According to the very Microsoft documentation of performance analysis, it is true that good troubleshooting requires good data.
This is not a technological manual but a guide on how to form a better relationship with your computer. Getting to know how to log and benchmark means you are no longer a mere user but a knowledgeable observer. You can also answer such questions as: "Is my slowdown occurring at the same time on a daily basis? or What did my PC last do the last thing before that freeze? With this kind of knowledge all things are different. It transforms frustration into recognition and chance troubleshooting into a procedural, gratifying undertaking. Now we can begin this voyage of becoming the actual expert on your own system.
Learning the Performance Monitor Ecosystem.
Before we begin to press buttons, we shall construct a basic mental model. Performance Monitor does not only measure speed. It hearkens to dozens of various system parts and converts their activity into a regular, numerical language which you can study.
Fundamental Elements: Objects, Counters, and Instances.
PerfMon is beautiful because of its logical structure. When using three straightforward words, the entire tool is available:
Objects:
These are the large systems within your PC. Consider them as characters of a story. Standard items are Processor, Memory, Physicaldisk (your hard drive or SSD), Network Interface, and Processes (the programs themselves which are in operation).
Counter:
The definite measurements of every thing. They give an account of what the object is doing. With the Memory object, one of the important counters is the Available MBytes which informs you of the amount of free RAM that you have at a given moment. In the case of the Physicaldisk object, the percentage of the time the drive is occupied can be known by using the physicaldisk object by a value of percent disk time.
Cases:
There are numerous copies of many objects. It is probably that your processor has many cores. An example allows you to trace a counter of a particular one. You may view "Processor% Processor Time" either of the Core 0, Core 1 or the instance named as the total which is an average of all of them.
This is a PerfMon grammar pattern Object:
Counter: Instance. You will see it everywhere and you know that it is your first step to understand communication with your tool is clear.
Real-Time Monitoring vs. Long-Term Data Logging:
Why They are both important.
PerfMon provides two primary data views each with its own needs.
Real-Time Monitoring (The Live Snapshot):
This is the graphics of the flashing updating line chart that you get the first thing you open PerfMon. It is wonderful to check something now. To add the percentage Processor Time counter in real time and see which process is spiking the fan, just add it the second after your fan is roaring. It is the response to What is happening now? The catch? You open the window and you lose the information. It does not know what caused an accident that occurred one hour ago.
Data Logging (The Time-Lapse Video):
This is where the DIY benchmarking is. Handing PerfMon the list of counters you want to monitor silently to a file, after every few seconds, or every few minutes or every hour--begone--is called logging. It is established by you once, and it collects a history. It is your treasure trove of this log file. It allows you to time travel and examine the precedents of a problem. Was the disk maxed out? Did memory slowly fill up? The logging provides you with the situation that live snapshots always lacks. It is the difference between seeing a picture of a traffic jam and viewing a video that will present how the traffic jam began.
Your First DIY Data Collector Set: Making a Data Collector Set.
PerfMon has the Data Collector Set which is used to enable automated logging. It is a stored program which writes: "Record this particular counter at this rate and save the information here. Let's create one together.
An Intensive, Step-by-Step Guide.
We are going to develop a simple log to record the overall well-being of your system. Follow these steps with me:
Open Performance Monitor:
Search performance monitor in your windows start menu search bar and open the application.
Find Data collector Sets:
under the left-hand menu, under data collector sets, expand. Right-Click on User Defined, and then New, and then Data Collector Set.
Name Your Project:
Name your project something distinct that you will remember such as myPC Baseline Log. Click Next, Select Create manually (Advanced).
Select What to Log:
Make sure that "Create data logs is selected, and you should choose the box of Performance counter. Click Next.
Add Counters - The Center of the Process:
Click Add... button. A new window opens. To a super starter set, add:
Add percentage Processor time:
Processor time. Select the "_Total" instance.
Based on the Memory object:
Sum up Available MBytes, and pages/sec.
Based on the Physical Disk object, Add PhysicalDisk, Disk Time and Avg. Disk Queue Length. Choose your primary drive (most often, C:) in the instances list.
Adding to the Network Interface object:
add Increasing the Instance selection of your active network connection is the addition of the object: Bytes Total/sec.
Set the Pace (Sample Interval):
This is used to set the frequency of taking data. To have a good starting point, then have this at 15 seconds. It provides details in great detail without gigantic files. Click Next.
Where to Save:
Select a location in a drive with work area. You can use a special folder (such as C:\PerfLogs.
End and Start:
Click Finish. Your new set is found under "User Defined." Right clicking it and selecting Start is necessary to log in. You'll see a green play icon. It has gone to rest without a murmur, and is collecting your data.
Choosing Counters with Purpose:
Increasing Metrics to Your Questions.
The counters that you select must be in direct relation to what you wish to learn.
For General "Is my PC healthy?" logging:
The set of starter is good.
In case you think there is a problem with memory:
Focus on Memory\Available Mbytes and Process Working set of particular applications. The typical symptom is a steadily deteriorating Available MBytes with time.
To examine disk slowness:
Examine PhysicalDisk. Avg. Disk sec/Read and Avg. Disk sec/Write. The values exceeding 20 milliseconds (0.020s) are usually indicative of a straining disk.
To check apps:
With Process object. Locate your instance of an app (such as chrome.exe) and add counters such as the percentage processor time.
Start simple. It is preferable to make 5 relevant counts correctly than 50 irrelevant counts that you do not comprehend.
Logs to Insight Data interpretation.
Gathering information is not the only thing. And the actual success is in making it make sense. PerfMon provides you with excellent analysis tools on your logs.
Contextualizing the Reports and Graphs.
After stopping your Data Collector Set, you can see its results. Select Reports and then User Defined and then [Name of your set]. PerfMon creates a summary of averages and highs/low.
To examine further and have a more interactive look, return to the Performance Monitor screen. The one resembling a database is the button to be clicked in the graph area, or Ctrl+L can be used. On the Source tab, change the default to 2: log files, the Click Add, and browse to your saved.blg file. At this point, you are able to add and delete counters to this historical data. You can zoom on a particular bad time and re-run it and see the interaction between the various resources. It is a good means to locate difficulty, ad hoc issues.
Determining Real-World Bottlenecks and Patterns.
Analysis is regarding the search of the number of relationships and numbers too high. The following are what you are going to seek in your data:
CPU Story:
When the value of the percentage of the processor time is consistent with more than 85-90, your processor is probably the issue. Determine what process occurs during those peaks.
Memory Story:
The amount of memory that should be available is comfortable. When it is repeatedly low (more than 10% of your total RAM) and Pages/sec is always high (more than 20) your system is being thrashed. This translates to it wasting extra effort in transfers of data between RAM and the slower page file in your disk drive which is a performance killer. Those who (or those who want to learn more) may detail the official Windows Sysinternals documentation provides advanced tools and in-depth commentary on these notions.
Space with a high percentage of Disk Time (close to 100%), and an average Disk Queue Length that is always greater than 2 per one disk, indicates a disk bottleneck. The disk simply can't keep up.
The "Aha!" Moment - Correlation:
It is the most informative. The time-synced log view can be used to determine whether several issues occurred simultaneously. Was the increase in Pages/sec (memory stress) at the same point in time when the percentage of Disk Time reached 100%? It was an easy to tell tale: a memory deficiency had necessitated the disk being overutilized and as a result, the system would slow down system wide.
Best Practices to Sustainable and Helpful Monitoring.
Logging is aimed at understanding and not additional work. There are a couple of easy routines that will make your monitoring handy and manageable.
Organizing to Clarity Not Chaos.
Develop various Data Collector Sets based on various jobs:
The Baseline Set:
The Baseline Set is a lightweight and long duration set (e.g. 30-second interval) that uses core counters to capture the normal state of your PC. Operate this over several days in the course of your normal working days in order to have your own point of reference.
The Investigation Set A high frequency (1-5 second interval) set which you initiate when you are actively attempting to catch a particular, repeatable problem.
Application Set:
This is a set used to profile a single program, which enables you to know how the program consumes its resources.
The Wise Management of Log Files.
Logs can be increased, but they are under your control.
Maximum Log folder:
Within your Data Collector Set Properties, as you are in the Directory Tab, you have the opportunity to establish a limit on the size of folder. This prevents logs consuming your drive space.
Use a Simple Cleanup Rule:
Under the same menu, there is an option to "Delete old data" when the limit is obtained. This forms a rolling log which never exceeds your set size.
Select a Rational Timeframe:
Trend analysis does not require a data point on a second-by-second basis. One minute or 30 seconds time is commonly ideal in recognizing daily trends as well as maintaining files small.
Smart Logs:
Conclusion
The art of using windows performance monitor does not entail mastery as a system administrator; it entails being a more assured companion to your system. Our initial friendly irritation was the basic annoyance of slowing unexpectedly and we ended up with a route to get rid of the guess work and knowledge. You have learned the logic language of PerfMon, how to configure the automated logging that records what is happening in your system and how to interpret the information in that log to extract the real answers.
This adventure transforms your level of association with technology. Being a victim of an unfamiliar machine turns into being a spectator with a set of instruments. You are also able to set what normal should be in your PC, diagnose issues using facts, and determine whether the changes you are implementing are beneficial. Start small. Write up that initial baseline log. Have a feel of the story it describes of an ordinary day. The self-confidence derived out of such knowledge is the actual objective. You are now prepared to not only make use of your computer, but also to get to know it.
Frequently Asked Questions
I'm not a tech expert. Am I too complex when it comes to Performance Monitor?
Not at all. Although it is a potent business instrument, its fundamental concepts can be easily acquired. This is the guide that came to clarify. Begin with the easy-steps to make a basic Data Collector Set. Only learn about 2-3 important counters at once e.g. CPU and Memory. You do not have to be aware of everything at once. Imagine you are learning some new hobby, you have to begin with the fundamentals and gradually increase the knowledge at your own pace.
What is the most effective frequency of sampling data to avoid large files?
This is a very useful question. A 15-second to 1-minute interval is the best value when it comes to long-term trend logging (such as a baseline that you're running over a week-long period). It provides a clear image of activity and files are very tiny. In cases where you want to find a particular and fast issue, you may temporarily set a time interval of 1 or 5 seconds as you search and reverse to your previous setting. Match the tool to the task.
Is it possible to examine my desktop computer logs on my laptop?
Yes, you can. The blg log files are portable. They can be copied to a second Windows computer, and then Performance Monitor can be opened in that computer and then the data source of the log files loaded and analyzed. It comes in handy when you wish to review logs of a strong desktop on a smaller computer, or when you simply wish to have your analysis detached to the machine under investigation.
What which counter is the one that I should monitor in case my PC is simply feeling slower?
A single counter is hardly enough to tell the whole story, so begin with a combination. Check first Memory Available MBytes. When it is very low, then you are probably short of RAM. Subsequently, examine Physicaldisk% Disk Time or Avg. Disk Queue Length. When your disk is continually busy at 100 percent, then that is a huge bottleneck. The two mentioned above, namely, running out of available memory, and having a disk filled to capacity, are the most frequent, on the one hand, general system slow-downs, of which you can observe and be quite sure.