https://github.com/bgeneto/windows-disk-trace-vis
A simple but useful web application designed to streamline the process of analyzing Event Trace Log (ETL) files, specifically focusing on Disk I/O activities
https://github.com/bgeneto/windows-disk-trace-vis
benchmarking performance-analysis ssd storage tracing windows
Last synced: 2 months ago
JSON representation
A simple but useful web application designed to streamline the process of analyzing Event Trace Log (ETL) files, specifically focusing on Disk I/O activities
- Host: GitHub
- URL: https://github.com/bgeneto/windows-disk-trace-vis
- Owner: bgeneto
- License: mit
- Created: 2023-08-21T23:10:01.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2026-02-05T00:35:12.000Z (6 months ago)
- Last Synced: 2026-02-05T11:40:51.644Z (6 months ago)
- Topics: benchmarking, performance-analysis, ssd, storage, tracing, windows
- Language: Python
- Homepage: https://windows-disk-trace-vis.streamlit.app/
- Size: 3.46 MB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Windows Disk Trace Visualizer
[Windows Disk Trace Visualizer](https://windows-disk-trace-vis.streamlit.app/) is a simple but useful web application designed to streamline the process of analyzing Event Trace Log (ETL) files,
specifically focusing on Disk I/O activities. This tool offers an intuitive and user-friendly interface to effortlessly navigate
through complex ETL data, providing insightful summaries and statistics for a thorough understanding of your disk performance requirements.
Key Features:
- **Data Analysis:** Uncover the most frequently requested data size, the percentage of random reading or writing, the average access time, and more.
- **Performance:** Identify the average throughput and IOPS for each disk, and compare the performance of sequential and random requests.
- **Filter Capability:** Filter data by process name or request size.
All this info is valuable for understanding where the performance of your disk/ssd matters most for your desired/traced workload/application.
# Usage
The webapp is hosted [here](https://windows-disk-trace-vis.streamlit.app). You will need my custom view
[profile](https://raw.githubusercontent.com/bgeneto/windows-disk-trace-vis/main/DiskIO.wpaProfile) to convert the .etl trace log file to a .csv file.
First, you need to record a trace of your 'Disk I/O activity' with WPRUI.exe (or via command line). You can trace your Windows boot process or a specific application/workload.
Then you need to convert the saved `.etl` file to a `.csv` file using the `wpaexporter.exe` tool that comes with [Windows ADK](https://learn.microsoft.com/en-us/windows-hardware/get-started/adk-install).
The `wpaexporter.exe` tool is typically located in the `C:\\Program Files (x86)\\Windows Kits\\10\\Windows Performance Toolkit` folder, an example usage follows:
```cmd
wpaexporter.exe -i boottrace.etl -profile DiskIO.wpaProfile -delimiter ;
```
You can also trace your boot process as follows:
- Open an elevated command prompt and run:
```cmd
wpr -addboot GeneralProfile.Light -filemode -recordtempto D:\\Temp
```
After this, the trace will start automatically at the early stage of the next (re)boot.
- The command syntax to save the boot trace (.etl file) is the following:
```cmd
wpr -stopboot D:\\Temp\\boottrace.etl
```
Now you can upload the (compressed) `.csv` file to this page. The page script — written in Python{python_svg}using Pandas{pandas_svg} — will analyze the single trace record using the provided
[profile](https://raw.githubusercontent.com/bgeneto/windows-disk-trace-vis/main/DiskIO.wpaProfile) and show you the results in tabular and graphical formats.
> **Note:** tested with WPR v10.0.25931.1000 on Windows 11 Pro version 22H2.
# Example Output
Here is an excerpt from the application's output, which demonstrates a typical result from our Windows 11 boot trace [example file](https://raw.githubusercontent.com/bgeneto/windows-disk-trace-vis/main/Disk_Usage_Counts_by_IO_Type_Priority.csv.bz2).
| | Value |
| ----------------------- | :------------ |
| Monitoring time | 52.25 seconds |
| Average access time | 83.93 µs |
| Read requests | 150704 |
| Write requests | 15948 |
| Total requests | 166652 |
| Percent READ | 90.43% |
| Percent WRITE | 9.57% |
| Percent RANDOM | 90.07% |
| Percent SEQUENTIAL | 9.93% |
| Read data size | 3.65 GB |
| Write data size | 0.29 GB |
| Total data size | 3.95 GB |
| Min. read request size | 0.5 KB |
| Avg. read request size | 25.4 KB |
| Max. read request size | 43132.5 KB |
| Min. write request size | 0.5 KB |
| Avg. write request size | 19.3 KB |
| Max. write request size | 1136 KB |
☝ Note the predominant number of 4KB random read requests (56%), followed by 16KB requests (18% read|30% write).
☝ Time spent in every request size.
☝ Note how much faster is an SSD vs HDD in terms of latency/access time.
☝ Random vs Sequential requests by IO Type.
# About WPR
Windows Performance Recorder ([WPR](https://learn.microsoft.com/en-us/windows-hardware/test/wpt/windows-performance-recorder))
is a performance recording tool that is based on Event Tracing for Windows ([ETW](https://learn.microsoft.com/en-us/windows/win32/etw/about-event-tracing)).
It records system and application events that you can analyze by parsing the generated Event Trace Log ([ETL](https://learn.microsoft.com/en-us/windows-hardware/drivers/devtest/trace-log)) file.
The ETL file format is a proprietary binary file format used by Microsoft Windows for collecting
and storing event traces generated by various components of the operating system. Since it is a poorly documented and
proprietary format, we need to use the [WPA Exporter](https://learn.microsoft.com/en-us/windows-hardware/test/wpt/exporter)
tool in order to convert the .etl file to a .csv file and then perform the desired trace analysis via csv file. Both WPR and WPA Exporter are part of the
Windows Assessment and Deployment Kit [Windows ADK](https://learn.microsoft.com/en-us/windows-hardware/get-started/adk-install).