Automating Image Pipelines with HCP Packer
summary
HashiCorp builds an automated image pipeline that uses HCP Packer channels as a cloud‑agnostic image registry and GitHub Actions to run Packer builds, spin up test infrastructure with Terraform, and promote images. The org runs ~70,000 CI jobs/month (≈16,000 compute hours) across Ubuntu, macOS, and Windows; builds create an iteration id per Packer run, an integration test job runs end‑to‑end against the baked image, and promotions move iterations to dev/staging and after 24 hours to production with the ability to revoke or roll back.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:26CI/CD at HashiCorp
- 1:34Actions Runner Fleet
- 2:26Self-Hosted Runner Fleet
- 3:04Self-Hosted GitHub Actions Runners
- 5:09Achieving Confidence
- 6:22Build & Deployment Pipeline
- 11:32What about HCP Packer?
- 12:13Some HCP Packer Terminology: Bucket
- 16:42Image Promotion Job
- 24:20Image Automation
- 25:16Resources
related talks
37mGolden Image Factory With HCP Packer & Terraform
56mHCP Packer Learn Lab: Build a Golden Image Pipeline
33mBuilding Automated Pipelines for Infrastructure Code with Terraform and Packer
29mDeploying CI/CD Infrastructure with Terraform and Packer
33mPacker: A Swiss Army Knife for Architectural Enablement
11mAutomating Cloud Security Compliance: From Patching to Machine Image Building and Auto Scheduling
52mBuilding Secure and Flexible Multi-Cloud Images with Multi-Boot Mode
30mA beginner's journey: Scaling self-hosted platform infrastructure with Nomad