Workflow Deep Dive

LinkedIn Draft — Workflow (2026-03-03) {{opener}} Terraform DAGs aren’t deterministic at scale — your abstractions are Terraform’s graph is great at parallelism, not safety. The sharp edges show up when the graph gets wide (mono-repos, shared modules, many resources). What usually bites later: Implicit dependencies hide ordering until a refactor turns into a surprise destroy/create chain. Fan-out graphs (100s of resources) explode apply time and make blast radius review basically impossible. depends_on “fixes” symptoms, then quietly couples modules and kills reuse. My default rule: If a module needs depends_on to be safe, the module boundary is leaking — redesign the interface, don’t paper over it. When I’m sanity-checking this, I usually do: Use terraform graph | dot -Tsvg > graph.svg and inspect fan-out + cycles before big refactors. Gate applies with policy (OPA/Conftest or Checkov) + a human review on any planned destroys. Deep dive (stable link): https://neeraja-portfolio-v1.verce