Sinatra CI/CD pipeline with Travis

This repository contains a simple Sinatra web application, the unit tests and a
CI/CD pipeline using Travis-CI.

Table of Contents

• The Application
  • Requirements
    • System
    • Gems
  • Usage
    • Vendor
    • Running
  • Environment
  • Database Schema
  • API
  • Tests
• The Pipeline
  • Steps
    • Test
    • Build
    • Publish

The Application

The web application tries to establish a RESTful API that interacts with a
PostgreSQL database. It simply create new “person” records on POST requests and
query by the id on GET requests.

Requirements

System

• Ruby 2.2.10 (or superior).

Gems

• sinatra
• rake
• rake-test

Usage

Vendor

To install the required libraries/3rd-party dependencies, use:

$ make install

This target will install bundler and, with bundler, install the necessary gems.
Alternatively, just use:

$ bundle install

Execution

To start the service, use:

$ make run

This target uses rack, so the service port will be binded, by default, to
9292. Alternatively, just use:

$ ruby src/app.py

To do it directly. In this case, the service port will be binded to 4567.

Environment

The following environment variables are used by the service:

Database Schema

The service expects an already created database with a table called users.
The table should have the following schema:

id        SERIAL
name      VARCHAR(255)
surname   VARCHAR(255)
age       INT

After connecting to the database:

• Create the database

CREATE DATABASE sinatra;
USE sinatra;

• Create the table users

CREATE TABLE users (id SERIAL, name VARCHAR(255), surname VARCHAR(255), age INT);

API

The application exposes three routes:

• GET /

  Returns a “Hello World“.

• GET /users/:id

  Connects to the database and query the users table by the id. The
  response should be similar to:

    {
        "name": "John",
        "surname": "Williams",
        "age": "86"
    }

• POST /users

  Connects to the database and inserts a new row into the users table. The
  payload should contains the keys name (string), surname (string) and
  age (int).

  E.g. (using the http tool):

  • http POST :4567 name="John" surname="Williams" age=86

Tests

To execute the unit tests, use:

$ make test

The Pipeline

The pipeline tries to accomplish a flow where, given a new release (a tag), a
new Docker image is generated and pushed to a private Docker registry (in AWS,
with ECR. It’s the “CD“ (continuous delivery) aspect of it. The
“CI“ (continuous integration) is achieved by the unit tests.

Steps

Test

• Dependency : N/A.
• Condition : push, pull request and tag
• Description : Installs the dependencies and execute the unit tests.

Build

• Dependency : test
• Condition : push on master or tag
• Description : Builds the Docker image.

Publish

• Dependency : build
• Condition : tag
• Description : Tags the builded image and pushes to the registry.