GitLab Services in CI/CD Pipelines

3 min read2 days ago

GitLab CI/CD allows users to run automated pipelines to build, test, and deploy applications. One of the key features of GitLab CI/CD is services, which enable running additional containers that support the main job execution. Services are commonly used to set up databases, caching layers, or other dependent services needed for tests and builds.

This blog post explains how GitLab services work with an example snippet and demonstrates practical use cases.

What Are GitLab Services?

In GitLab CI/CD, services are additional Docker containers that run alongside the primary job container. These containers provide dependencies such as databases, configuration servers, or message queues.

Services are especially useful when:

Running integration tests that require external services.
Setting up dependencies like databases, config servers, or caching layers.
Ensuring test environments closely resemble production setups.

Example: Using a Spring Cloud Config Server as a Service

Let’s break down the example GitLab CI/CD configuration snippet you provided:

.config-vars:
  services:
    - name: hyness/spring-cloud-config-server:4.1-jre17
      alias: localhost  
  before_script:
    - set -o allexport
    - source .env
    - set +o allexport
    - echo "Environment variables successfully set"  
    - echo "Waiting for config server to be ready..."
    - timeout 30 bash -c 'until curl -s "http://localhost:8888/actuator/health" | grep UP; do sleep 5; done' || echo "Service not reachable"
    - export SPRING_CLOUD_CONFIG_URI=http://localhost:8888/

Defining Services

services:
  - name: hyness/spring-cloud-config-server:4.1-jre17
    alias: localhost

Here, the hyness/spring-cloud-config-server:4.1-jre17 image is used as a service.
The alias localhost allows referencing this service using http://localhost:8888/.

Setting Environment Variables

before_script:
  - set -o allexport
  - source .env
  - set +o allexport

This snippet loads environment variables from an .env file, making them available in the job.

2. Waiting for the Service to Start

- echo "Waiting for config server to be ready..."
- timeout 30 bash -c 'until curl -s "http://localhost:8888/actuator/health" | grep UP; do sleep 5; done' || echo "Service not reachable"

The script continuously checks whether the config server is running. If the server is UP, the job continues; otherwise, it prints a failure message after 30 seconds.

3. Exporting the Config Server URL

- export SPRING_CLOUD_CONFIG_URI=http://localhost:8888/

This ensures that the application can access the configuration server by setting the appropriate environment variable.

4. Running Multiple Services

If your application depends on multiple services, you can define them under services:

services:
  - name: postgres:13
    alias: db
  - name: redis:latest
    alias: cache

This ensures that both PostgreSQL and Redis are available for your pipeline.

5. Using a Custom Docker Network

To improve communication between services, you can define a custom Docker network:

variables:
  DOCKER_DRIVER: overlay
services:
  - name: mysql:latest
    alias: mysql_db
    command: ["--default-authentication-plugin=mysql_native_password"]
  - name: rabbitmq:3-management
    alias: rabbit
    variables:
      RABBITMQ_DEFAULT_USER: user
      RABBITMQ_DEFAULT_PASS: password

6. Running Integration Tests with Services

If you need to run integration tests with services, use needs to ensure proper sequencing:

stages:
  - build
  - test

build-job:
  stage: build
  script:
    - echo "Building application..."integration-tests:
  stage: test
  services:
    - name: mongo:latest
      alias: mongo_db
  script:
    - echo "Running tests against MongoDB..."
  needs: [build-job]

Benefits of Using Services in GitLab CI/CD

Isolation: Each service runs in its own container, ensuring a clean test environment.
Reproducibility: CI/CD pipelines remain consistent since services are always started with predefined configurations.
Ease of Setup: Services eliminate the need for complex installation scripts by pulling ready-to-use Docker images.

Conclusion

GitLab services are an essential tool in CI/CD pipelines, enabling dependency management with minimal effort. Whether integrating a configuration server, database, or caching layer, services ensure that your CI/CD environment closely mirrors production. By leveraging GitLab services, teams can build more robust and reliable pipelines.