WAT.tf
summary
Terraform gets praised but also breaks expectations: using lists to DRY S3 bucket creation leads to one delete planning to destroy all buckets; modules do not behave like classes so the dependency graph doesn’t guarantee order and modules are not first-class citizens; running terraform on every deploy increases chances of unexpected adds/destroys and leaves orphan resources still serving traffic but unmanaged. Practical fixes: refactor to per-resource state, add tests and locking, or use the cloud SDK and real code for tricky logic.
AI-written summary · about
related talks
35mTerraform Practices - The Good, the Bad and the Ugly
35mMisadventures with Terraform
41mTerraform Practices - The Good, the Bad, and the Ugly
30mTerraform Practices to Enable Infrastructure Scaling
29mTerraform Practices - The Good, the Bad, and the Ugly
31mTerraform Practices to Enable Infrastructure Scaling
39mTerraform Practices to Enable Infrastructure Scaling
30mTerraform in Depth QA and Fireside Chat