flowtest

FlowTest

Declarative backend integration testing. Write YAML, test everything.

FlowTest reads YAML flow files and executes backend test sequences — API calls, database operations, Kafka events, Redis validation, and shell scripts — with variable passing, conditional execution, and expression-based assertions.

In 60 seconds

# 1. Install
curl -sSfL https://raw.githubusercontent.com/rakrisi/flowtest/main/install.sh | sh

# 2. Create a flow (no infrastructure required)
cat > hello.yaml << 'EOF'
name: Hello FlowTest
steps:
  - name: Script works
    script:
      run: echo "FlowTest is working!"
      assert_exit: 0
    assert:
      - expr: "stdout == 'FlowTest is working!'"
EOF

# 3. Run it
flowtest run hello.yaml

  ✓ Script works  (12ms)

  Passed 1/1 steps

Once your API/DB is running, scaffold a real project:

mkdir my-tests && cd my-tests
flowtest init          # interactive: picks only the services you use
flowtest run flows/example.yaml

Features

• Multi-database support — PostgreSQL, MySQL, MongoDB, SQLite (auto-detected from DSN)
• HTTP testing — GET/POST/PUT/PATCH/DELETE with headers, JSON body, response assertions
• Kafka validation — Produce messages and search/consume with field matching
• Redis operations — GET, SET, DEL, EXISTS, KEYS, TTL, HGETALL
• Shell scripts — Execute commands with environment injection and exit code assertions
• Variable chaining — Pass data between steps with ${var} and dot notation
• Expression assertions — Powered by expr-lang
• Setup/cleanup lifecycle — Seed databases before, clean up after (always runs)
• Retry with backoff — Linear or exponential retry for flaky steps
• Dry-run mode — Preview what would execute without running anything
• JSON output — Machine-readable results for CI/CD pipelines
• Profiles — Switch between environments (dev, staging, prod)

Quick Start

Install

macOS / Linux / Windows (recommended):

curl -sSfL https://raw.githubusercontent.com/rakrisi/flowtest/main/install.sh | sh

The script auto-detects your OS and architecture, downloads the correct binary from the latest release, and installs it. It falls back to go install if the archive is unavailable.

Go install:

go install github.com/rakrisi/flowtest/cmd/flowtest@latest

Build from source:

git clone https://github.com/rakrisi/flowtest.git
cd flowtest
go build -o flowtest ./cmd/flowtest/

Initialize

flowtest init

init is interactive — it asks which services you use (HTTP, PostgreSQL, MySQL, MongoDB, Redis, Kafka) and generates a minimal config with only what you need. Pass --no-interactive to skip prompts (defaults to HTTP only).

This creates:

• flowtest.yaml — connection strings for your selected services
• flows/example.yaml — a working example flow for those services

Configure

Edit flowtest.yaml with your connection strings:

env:
  api_base: http://localhost:8000

  databases:
    db: postgres://user:pass@localhost:5432/myapp
    # mysql_db: mysql://user:pass@localhost:3306/myapp
    # mongo_db: mongodb://localhost:27017/myapp
    # local: sqlite://./test.db

  # kafka_brokers: localhost:9092
  # redis: redis://localhost:6379

Database drivers are auto-detected from DSN schemes — no extra configuration needed.

Write a Flow

name: User API Test
description: Verify user CRUD operations

setup:
  - seed:
      table: users
      data: {email: test@example.com, name: Test User}

steps:
  - name: Create user
    api:
      method: POST
      url: /users
      body:
        email: new@example.com
        name: New User
    assert:
      - expr: "response.status == 201"
      - expr: "response.body.id > 0"
    save:
      user_id: response.body.id

  - name: Verify in database
    db:
      query: "SELECT * FROM users WHERE id = $1"
      params: [${user_id}]
    assert:
      - expr: "row_count == 1"
      - expr: "rows[0].email == 'new@example.com'"

  - name: Check Kafka event
    kafka:
      topic: user-events
      match: {user_id: ${user_id}}
      timeout: 5s
    assert:
      - expr: "message.payload.action == 'created'"

  - name: Verify cache
    redis:
      action: get
      key: "user:${user_id}"
    assert:
      - expr: "exists == true"

cleanup:
  - db:
      query: "DELETE FROM users WHERE email = 'new@example.com'"

Run

# Execute a flow
flowtest run flows/user-test.yaml

# Verbose mode — see requests, queries, assertions
flowtest run flows/user-test.yaml -v

# Dry-run — see plan without executing
flowtest run flows/user-test.yaml --dry-run

# Stop on first failure
flowtest run flows/user-test.yaml --fail-fast

# Inject variables
flowtest run flows/user-test.yaml --var token=abc123

# Use a profile
flowtest run flows/user-test.yaml -p staging

# JSON output for CI
flowtest run flows/user-test.yaml --output-json

CLI Reference

Command	Description
flowtest run <file>	Execute a flow
flowtest validate <file>	Parse and validate without executing
flowtest list [dir]	List all flow files
flowtest init	Scaffold a new project

Run Flags

Flag	Description
-p, --profile	Environment profile from flowtest.yaml
-v, --verbose	Show request/response details, queries, assertions
--fail-fast	Stop on first failure
--output-json	Output results as JSON
--html-report FILE	Generate HTML report (e.g., --html-report report.html)
--from-json	Load flow from JSON file
--var key=value	Set initial variables (repeatable)
--step N	Start from step N
--step-name	Start from the named step
--dry-run	Preview execution plan

HTML Reports

Generate comprehensive HTML reports with Tailwind CSS styling:

# Single flow with HTML report
flowtest run flows/api-test.yaml --html-report report.html

# Verbose mode + HTML report
flowtest run flows/integration.yaml -v --html-report results.html

Reports include:

• ✅ Pass/fail/skip status for each step
• ⏱️ Execution duration and timestamps
• 📊 Assertions with actual vs expected values
• 🔍 Request/response details (verbose mode)
• 📈 Summary statistics and step breakdown
• 🎨 Beautiful Tailwind CSS design

Tip: Open the HTML file in your browser to view the report. Perfect for CI/CD pipelines — commit reports to your repo or upload to artifacts.

Supported Drivers

HTTP

- name: Create item
  api:
    method: POST
    url: /items
    headers:
      Authorization: "Bearer ${token}"
    body:
      name: Test Item
      price: 29.99
    timeout: 10s
  assert:
    - expr: "response.status == 201"
    - expr: "response.body.name == 'Test Item'"
  save:
    item_id: response.body.id

SQL Databases (PostgreSQL, MySQL, SQLite)

# Query
- name: Verify item exists
  db:
    query: "SELECT * FROM items WHERE id = $1"
    params: [${item_id}]
  assert:
    - expr: "row_count == 1"
    - expr: "rows[0].name == 'Test Item'"

# Use a different database
- name: Check MySQL
  mysql_db:
    query: "SELECT COUNT(*) as total FROM orders WHERE user_id = ?"
    params: [${user_id}]
  assert:
    - expr: "rows[0].total > 0"

Seed in setup:

setup:
  - seed:
      target: db          # optional if only one database
      table: items
      data: {name: Seed Item, price: 10.0}

MongoDB

- name: Insert document
  mongo_db:
    collection: users
    operation: insertOne
    document:
      name: Test User
      email: test@example.com
  save:
    doc_id: rows[0]._id

- name: Find documents
  mongo_db:
    collection: users
    operation: find
    filter: {email: test@example.com}
  assert:
    - expr: "row_count >= 1"

- name: Update document
  mongo_db:
    collection: users
    operation: updateOne
    filter: {email: test@example.com}
    update:
      $set: {status: active}
  assert:
    - expr: "modified_count == 1"

Supported operations: find, findOne, insertOne, insertMany, updateOne, deleteOne, deleteMany, countDocuments.

Kafka

# Produce
- name: Send event
  kafka:
    action: produce
    topic: events
    key: order-123
    message: {action: created, order_id: 123}
    headers:
      event-type: order.created

# Consume/search
- name: Find event
  kafka:
    topic: events
    match: {order_id: 123}
    timeout: 10s
  assert:
    - expr: "message.payload.action == 'created'"

Redis

- name: Set value
  redis:
    action: set
    key: "session:${user_id}"
    value: {logged_in: true}
    ttl: 1h

- name: Check value
  redis:
    action: get
    key: "session:${user_id}"
  assert:
    - expr: "exists == true"
    - expr: "value.logged_in == true"

- name: Check TTL
  redis:
    action: ttl
    key: "session:${user_id}"
  assert:
    - expr: "expires == true"
    - expr: "ttl > 3500"

Actions: get, set, del, exists, keys, ttl, hgetall.

Shell

- name: Run script
  script:
    lang: bash
    run: |
      curl -s http://localhost:8000/health | jq -r '.status'
    assert_exit: 0
    timeout: 10s
  assert:
    - expr: "stdout == 'ok'"

Step Features

Conditional Execution

- name: Only if feature enabled
  when: "feature_flag == true"
  api:
    method: GET
    url: /feature-endpoint

Retry

- name: Wait for async processing
  retry:
    times: 5
    interval: 2s
    backoff: exponential
  api:
    method: GET
    url: /status/${job_id}
  assert:
    - expr: "response.body.status == 'completed'"

Delay

- name: Wait before checking
  delay: 2s
  redis:
    action: get
    key: "result:${job_id}"

Variable Chaining

Steps pass data to subsequent steps via save:

- name: Create order
  api:
    method: POST
    url: /orders
    body: {item: Widget}
  save:
    order_id: response.body.id

- name: Check order
  api:
    method: GET
    url: /orders/${order_id}
  assert:
    - expr: "response.body.id == order_id"

Multi-Database Configuration

Configure multiple databases by name:

# flowtest.yaml
env:
  databases:
    postgres_db: postgres://user:pass@localhost:5432/app
    mysql_db: mysql://user:pass@localhost:3306/app
    mongo_db: mongodb://localhost:27017/app
    local: sqlite://./test.db

Reference them in flows by their configured name:

steps:
  - name: Query Postgres
    postgres_db:
      query: "SELECT * FROM users"

  - name: Query MySQL
    mysql_db:
      query: "SELECT * FROM orders"

  - name: Query MongoDB
    mongo_db:
      collection: logs
      operation: find
      filter: {level: error}

Profiles

Define environment profiles for different targets:

# flowtest.yaml
env:
  api_base: http://localhost:8000
  databases:
    db: postgres://user:pass@localhost:5432/app

profiles:
  staging:
    api_base: https://staging.api.example.com
    databases:
      db: postgres://user:pass@staging-db:5432/app

flowtest run flows/test.yaml -p staging

Try the full demo

The testproject/ directory is a complete, runnable reference — a Go API server wired to Postgres, MySQL, MongoDB, Redis, and Kafka, with 9 flow files covering every driver.

cd testproject
./run-tests.sh

Requires Docker. The script:

1. Builds the FlowTest binary and a test API server
2. Starts all services via Docker Compose (Postgres, MySQL, MongoDB, Redis, Kafka)
3. Validates every flow file
4. Runs all flows with verbose output and generates HTML reports in test-reports/

Use it as a copy-paste reference for your own flows.

Development

# Run unit tests
go test ./...

# Run with verbose
go test ./... -v

# Build
go build -o flowtest ./cmd/flowtest/

# Vet
go vet ./...

License

MIT