Problem:
Currently, benchmark results are saved as JSON files in the wrk-results folder. These are committed to the repository every time someone runs their local benchmarks. This pollutes the repo history and makes collaboration harder, especially when users are testing locally or tuning implementations.
Proposed Solution:
In the future, develop a small web dashboard or static site to display and compare the results across all implementations and frameworks. The dashboard should:
- Load results dynamically (e.g., from a separate branch, external storage, or API)
- Show response times, throughput, and error rates side by side
- Visualize historical performance changes (optional)
Possible Implementation Ideas:
- Host it via GitHub Pages or a subdomain on my personal domain
- Generate charts using a tool like Chart.js, Recharts, D3 or Matplotlib
- Automatically update from CI pipeline (optional stretch goal)
- Keep raw JSON results but exclude them from Git commits (e.g., add to .gitignore or store externally)
Why this matters:
It encourages more accurate benchmarking, prevents accidental result commits, and improves accessibility and collaboration around performance data
Problem:
Currently, benchmark results are saved as JSON files in the wrk-results folder. These are committed to the repository every time someone runs their local benchmarks. This pollutes the repo history and makes collaboration harder, especially when users are testing locally or tuning implementations.
Proposed Solution:
In the future, develop a small web dashboard or static site to display and compare the results across all implementations and frameworks. The dashboard should:
Possible Implementation Ideas:
Why this matters:
It encourages more accurate benchmarking, prevents accidental result commits, and improves accessibility and collaboration around performance data