Implementation Guidelines

$ Production

# Tuned

* Engine

Must load files from disk on every request. No in-memory caching, no memory-mapped files, no pre-loaded file buffers.

May cache files in memory at startup, use memory-mapped files, pre-rendered response headers, or any caching strategy.

No specific rules.

Serves 20 static files of various types and sizes over HTTP/1.1, simulating a realistic page load with diverse file types and sizes.

Connections: 4,096, 16,384

Workload

The load generator (gcannon) requests 20 static files in a round-robin pattern using raw HTTP/1.1 request templates:

CSS (5 files, 1.2-12 KB): reset.css, layout.css, theme.css, components.css, utilities.css
JavaScript (5 files, 3.2-35 KB): analytics.js, helpers.js, app.js, vendor.js, router.js
HTML (2 files, 1.1-1.5 KB): header.html, footer.html
Fonts (2 files, 32-38 KB): regular.woff2, bold.woff2
SVG (2 files, 4.5-8 KB): logo.svg, icon-sprite.svg
Images (3 files, 18-85 KB): hero.webp, thumb1.webp, thumb2.webp
JSON (1 file, 0.9 KB): manifest.json

Total payload: ~325 KB across 20 files.

GET /static/reset.css HTTP/1.1
Host: localhost:8080

HTTP/1.1 200 OK
Content-Type: text/css

(file contents)

GET /static/app.js HTTP/1.1
Host: localhost:8080

HTTP/1.1 200 OK
Content-Type: application/javascript

(file contents)

Parameter	Value
Endpoint	20 URIs under `/static/*`
Connections	4,096, 16,384
Pipeline	1
Duration	5s
Runs	3 (best taken)
Load generator	gcannon with `--raw` (multi-template)