Continuous Deployment with GitLab

Summary

The guide shows how to set up continuous deployment from GitLab CI/CD to a Fly application, using the small Go hellofly example. GitLab supplies the repository contents to a Runner defined by the golang:1.13 image, while a before_script installs curl and downloads flyctl through Fly’s installation script. A deploy job then runs flyctl deploy, which builds and deploys the application from a committed fly.toml file. Authentication is configured by obtaining a token with flyctl auth token and storing it as a protected, masked FLY_API_TOKEN CI/CD variable in the repository settings. After .gitlab-ci.yml is committed and pushed, GitLab starts the deployment job, while the guide notes that more advanced pipeline structures and triggers remain available.

Takeaways

1. 1

   GitLab automatically copies repository contents into the configured Runner, satisfying the application-source requirement for the deployment process.

2. 2

   The configuration installs curl and uses Fly’s installation script to obtain a runnable flyctl before the deployment job runs.

3. 3

   The API token is stored in GitLab’s CI/CD variables as FLY_API_TOKEN with Protected and Masked enabled to prevent exposure in logs.