How Global Discovery tribe approaches Infrastructure Changes at scale
summary
Claims a reproducible Terraform workflow that scales from a single user to dozens/hundreds by generating module-backed configs and centralizing plans and state. Uses GCP/GKE, remote state in GCS buckets, Terragrunt modules for reuse, a dockerized dh-infra CLI (Python) that renders templates and emits tf files, and Atlantis for PR plan/apply automation. Manages multiple deployments per project, supports cascading module releases, powers 50 million+ searches daily and is deployed successfully in production.
AI-written summary · about
chapters (9)
- 0:00Introduction and background
- 1:39Terraform at small scale
- 2:45Scaling Terraform workflow
- 3:51Tech stack overview
- 4:54In-house infrastructure tools
- 6:20Template and module management
- 7:46Usage in search tribe
- 11:13Centralized deployment with Atlantis
- 13:39Workflow demonstration
related talks
27mCase Study: Re-Thinking Our Infrastructure Tooling
29mInfinity and beyond: Scaling Terraform for the global enterprise
39mTerraform Practices to Enable Infrastructure Scaling
31mTerraform Practices to Enable Infrastructure Scaling
06mIGNITE: Terraform Practice to Enable Infrastructure Scaling
30mTerraform Practices to Enable Infrastructure Scaling
35mMisadventures with Terraform
24mTerraform Best Practices