How to build solid CI-CD pipeline
summary
Pipeline prescribes packaging every service as a Docker image built on an isolated builder (never on production), running tests and static analysis in CI, pushing tagged images to a registry, then running a separate deploy pipeline that executes an Ansible playbook inside Docker on an AWS-bound runner which pulls images and starts containers. Development uses Docker Compose for parity and integration tests live in an infrastructure repo. Proven specifics: Docker images built off isolated build server, GitLab CI runners (per-project) plus a dedicated AWS deploy runner, Ansible playbook packaged in Docker for deployment.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:15Manual deployment problems
- 6:49Automation solution
- 9:17Docker in CI/CD
- 12:27Auto Deploy: Docker
- 16:01Auto Deploy: Ansible
- 21:23Auto Deploy: GitLab CI
- 25:59Repository structure
- 33:52Releases and rollbacks
- 34:12Q&A
related talks
23mOrchestration and CI/CD
42mHow Containerized Pipelines Can Boost Your CI/CD
25mIncrease App Confidence Using CI/CD and Infrastructure as Code
15mDocker image testing in GitLab CI
46mFrom Dev to Prod with GitLab CI
01h 00mVintage CI/CD. How and Where to Start
06mBuilding Continuous Delivery Pipelines
26mOptimizing Your CI Pipelines