Loadzilla

Fast, zero-config HTTP load testing — run as CLI or open the browser UI in one command.
Optional CSV import to parameterize field and start firing custom requests.
You choose concurrent threads.
You choose requests per second.
You choose duration.
Live progress and nstant results

---

Install

go install github.com/lawi22/loadzilla@latest

Or download a pre-built binary from Releases.

---

Web UI

loadzilla serve          # opens on http://localhost:7777
loadzilla serve -p 9000  # custom port

Configure your test with checkboxes — only set what you need, everything else uses a sensible default.

Upload a CSV to parameterise any field. Column names become {$COLUMN} tokens you can click into the URL, headers, or body. Loadzilla cycles through rows round-robin across all workers.

Hit Run Test and watch results stream in live. The progress bar ticks every second; the full latency breakdown appears the moment the test finishes.

---

CLI

The CLI is unchanged and needs no server running.

# Simple GET, all defaults (10 connections, 10s)
loadzilla -u https://httpbin.org/get

# 20 connections, 200 rps cap, 30 seconds
loadzilla -u https://api.example.com/users -c 20 -r 200 -d 30s

# POST with headers and body
loadzilla -u https://api.example.com/users \
  -m POST \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -b '{"name":"alice","role":"admin"}' \
  -c 10 -r 50 -d 30s

# Read body from file
loadzilla -u https://api.example.com/import -m POST \
  -H "Content-Type: application/json" \
  --body-file ./payload.json \
  -c 20 -d 1m

Flags

Flag	Short	Default	Description
--url	-u	required	Target URL
--method	-m	GET	HTTP method
--header	-H	—	Key: Value header — repeatable
--body	-b	—	Request body as a raw string
--body-file	—	—	Read request body from a file
--connections	-c	10	Concurrent workers (goroutines)
--rps	-r	0	Max requests/sec — 0 = unlimited
--duration	-d	10s	Test duration (e.g. 30s, 2m)

Press Ctrl+C at any time to stop early — a full summary still prints.

CLI output

[████████████░░░░░░░░]  62% | 1,240 req | 97.8 rps | Errors: 3

Loadzilla Results
═════════════════════════════════════════════
 URL:          https://api.example.com/users
 Method:       POST
 Duration:     30s
 Connections:  20 | RPS Target: 200

 Requests
 ─────────────────────────────────────────────
 Total:        5,842
 Success:      5,839  (99.9%)
 Errors:       3  (0.1%)
 Throughput:   194.7 req/s

 Latency
 ─────────────────────────────────────────────
  Min:     12.4ms    P50:     38.2ms
  Avg:     41.1ms    P95:     76.4ms
  Max:    892.1ms    P99:    124.8ms
                     P999:   341.2ms

 Status Codes
 ─────────────────────────────────────────────
  200: 5,839
  500: 3
═════════════════════════════════════════════

---

How it works

• N goroutines run concurrently, each sending requests in a tight loop
• A shared token bucket (golang.org/x/time/rate) enforces the RPS cap globally
• net/http.Transport is tuned with MaxConnsPerHost matching --connections — no silent pooling surprises
• Latency is recorded into a mutex-protected HDR histogram (µs precision) for accurate percentiles at any volume
• Ctrl+C (CLI) or closing the tab (web UI) cancels the context — in-flight requests drain cleanly before the summary prints

---

License

MIT