Integration Testing C++ with Docker: Advanced Level

Integration Testing C++ with Docker: Advanced Level

The development of modern C++ applications is increasingly moving towards the paradigm of microservices and containerization. Integration Testing becomes a critical stage, allowing verification of component interactions in conditions as close to production as possible. Docker provides an ideal environment for isolated and reproducible testing. In this article, we will explore advanced techniques for integration testing C++ using Docker, including container setup, writing tests with Google Test, and integration into CI/CD pipelines.

1. Why Docker is the Ideal Tool for C++ Integration Testing?

Traditional unit testing checks individual functions or classes but does not guarantee the correct operation of the system as a whole. Integration tests verify connections: databases, network protocols, file systems, external APIs. This is where Docker becomes indispensable:

• Reproducibility: A Docker image contains an exact copy of the environment (compiler, libraries, environment variables). A test that passes locally will pass on the CI server and for another developer.
• Isolation: Each test runs in a clean container. There is no "contamination" from global OS settings or residual data from previous tests.
• Parallelism: You can run multiple test containers simultaneously, saving time in CI/CD.
• Dependency Management: Docker Compose allows you to spin up an entire infrastructure: PostgreSQL, Redis, Kafka — in seconds.

2. Setting Up a Test Environment with Docker Compose

Let's consider a typical scenario: a C++ service that reads data from PostgreSQL and sends it to RabbitMQ. We need to test the integration of these three components.

Let's create docker-compose.test.yml:

version: '3.8'services:  postgres_test:    image: postgres:15-alpine    environment:      POSTGRES_DB: testdb      POSTGRES_USER: testuser      POSTGRES_PASSWORD: testpass    ports:      - "5432:5432"    healthcheck:      test: ["CMD-SHELL", "pg_isready -U testuser -d testdb"]      interval: 5s      timeout: 5s      retries: 5

  rabbitmq_test:    image: rabbitmq:3-management-alpine    environment:      RABBITMQ_DEFAULT_USER: testuser      RABBITMQ_DEFAULT_PASS: testpass    ports:      - "5672:5672"    healthcheck:      test: ["CMD", "rabbitmq-diagnostics", "check_port_connectivity"]      interval: 5s      timeout: 5s      retries: 5

  test_runner:    build:      context: .      dockerfile: Dockerfile.test    depends_on:      postgres_test:        condition: service_healthy      rabbitmq_test:        condition: service_healthy    environment:      DB_HOST: postgres_test      DB_PORT: 5432      DB_NAME: testdb      DB_USER: testuser      DB_PASS: testpass      RMQ_HOST: rabbitmq_test      RMQ_PORT: 5672

The key point is healthcheck. We do not run tests until the database and message broker are ready to operate. This prevents "race conditions" in tests.

3. Writing Integration Tests in C++ with Google Test

For demonstration, we will use Google Test (gtest) — the de facto standard for C++ testing. Let's write a test that verifies inserting data into PostgreSQL and reading it via RabbitMQ.

First, let's create a simple wrapper for working with RabbitMQ (simplified):

// rabbitmq_client.h#pragma once#include <string>#include <amqpcpp.h>

class RabbitMQClient {public:    RabbitMQClient(const std::string& host, int port,                   const std::string& user, const std::string& pass);    bool sendMessage(const std::string& queue, const