veritas-kanban/load-tests

Load Tests

Performance and load tests for the Veritas Kanban API using k6.

Prerequisites

Install k6

macOS (Homebrew):

brew install k6

Other platforms: See k6 installation docs.

Verify installation

k6 version

Start the server

# From repo root
pnpm dev

Server should be running at http://localhost:3001.

Test Scenarios

Script	VUs	Duration	Description
smoke.js	1	1 iteration	CRUD lifecycle — create, read, update, delete
read-load.js	50	30s	Read-heavy: list + detail endpoints
write-load.js	20	30s	Write-heavy: create, update, delete
mixed-load.js	0→30	60s	70% reads / 30% writes with ramp-up
ws-stress.js	25	30s	WebSocket connection stress

Running Tests

Quick smoke test

pnpm test:load:smoke
# or directly:
k6 run load-tests/k6/smoke.js

Run all load tests

pnpm test:load

Scheduled CI runs the k6 smoke profile weekly and can run the full profile by manual dispatch. See docs/testing/scheduled-qa-gates.md for the scheduled/manual gate, artifact, and threshold contract.

Run a specific scenario

k6 run load-tests/k6/read-load.js
k6 run load-tests/k6/write-load.js
k6 run load-tests/k6/mixed-load.js
k6 run load-tests/k6/ws-stress.js

Override configuration via environment variables

# Custom server URL
k6 run -e BASE_URL=http://localhost:4000 load-tests/k6/smoke.js

# Custom API key
k6 run -e API_KEY=my-secret-key load-tests/k6/read-load.js

# Custom WebSocket URL
k6 run -e WS_URL=ws://localhost:4000/ws load-tests/k6/ws-stress.js

Adjust VUs / duration on the fly

# Override VU count and duration
k6 run --vus 100 --duration 60s load-tests/k6/read-load.js

Interpreting Results

After a run, k6 prints a summary like:

     ✓ list → 200
     ✓ detail → 200

     checks.....................: 100.00% ✓ 4820  ✗ 0
     http_req_duration..........: avg=12.3ms  min=2.1ms  med=10.5ms  max=95.2ms  p(90)=22.1ms  p(95)=31.4ms
     http_reqs..................: 4820    160.6/s
     errors.....................: 0.00%   ✓ 0     ✗ 4820

Key metrics

Metric	What it means	Target
http_req_duration p(95)	95th percentile response time	<200ms reads, <500ms writes
errors	Percentage of failed checks	<1%
http_reqs	Total requests / requests per second	Higher = better throughput
checks	Assertion pass rate	100% for smoke, >99% for load
ws_errors	WebSocket connection failure rate	<5%
ws_messages_received	Total WS messages received	>0 per connection

Thresholds

Tests define built-in thresholds. If a threshold is breached, k6 exits with code 99:

• Read-heavy: p95 < 200ms, errors < 1%
• Write-heavy: p95 < 500ms, errors < 1%
• Mixed: p95 < 500ms, errors < 1%
• WebSocket: connection errors < 5%

Export results (optional)

# JSON output
k6 run --out json=results.json load-tests/k6/read-load.js

# CSV output
k6 run --out csv=results.csv load-tests/k6/read-load.js

File Structure

load-tests/
├── README.md          # This file
├── config.js          # Shared config (base URL, headers, helpers)
└── k6/
    ├── smoke.js       # Scenario 1: CRUD smoke test
    ├── read-load.js   # Scenario 2: Read-heavy load
    ├── write-load.js  # Scenario 3: Write-heavy load
    ├── mixed-load.js  # Scenario 4: Mixed workload
    └── ws-stress.js   # Scenario 5: WebSocket stress