HashiCorp Certified: Terraform Associate (004) study guide

How to use this guide

What to focus on:

• Why state exists, and what breaks without it
• Remote backends, state locking, and why local state fails a team
• Moving and importing resources with terraform state mv and import blocks
• Reading state safely, and why secrets in state matter

What to focus on:

• The write, init, plan and apply cycle, and what each step actually does
• Reading a plan and predicting create, update, replace and destroy
• terraform fmt and validate in a real workflow
• Targeted operations and -replace (which supersedes the older taint)

What to focus on:

• Variables, locals, outputs, and precedence when a value is set twice
• Expressions, functions, and dynamic blocks
• count versus for_each, and why for_each is usually safer
• Resource dependencies, implicit and explicit

What to focus on:

• Writing and calling modules, with inputs and outputs
• Sourcing from the public registry and from private sources
• Version constraints and why pinning matters
• Structuring a repository so modules stay reusable

What to focus on:

• What IaC solves: repeatability, review, and drift
• Declarative versus imperative provisioning
• Terraform's multi-cloud, provider-based model
• Idempotency, and what it means for a rerun

What to focus on:

• Providers, provider versions, and the dependency lock file
• Installing Terraform and the purpose of terraform init
• Resources versus data sources
• Provisioners, and why they are the last resort

What to focus on:

• Detecting and reconciling drift
• Upgrading providers and Terraform versions safely
• Refactoring with moved blocks instead of destroy-and-recreate
• Managing resource lifecycle with prevent_destroy and ignore_changes

What to focus on:

• Workspaces, and how remote runs differ from local ones
• Remote state sharing across workspaces
• Policy as code with Sentinel, and the private module registry
• VCS-driven workflows and team access

Common mistakes

• ✕Studying from 002 or 003 material. The retired exams teach commands and patterns that 004 no longer treats as correct, taint being the clearest example.
• ✕Reaching for count when the scenario describes resources that get added and removed. Changing a count re-indexes everything after it; for_each does not.
• ✕Underestimating state. It is where most real Terraform pain lives, and the exam reflects that.
• ✕Confusing the change symbols in a plan. Create, update in place, destroy and replace each have very different consequences.
• ✕Treating provisioners as a normal tool. HashiCorp documents them as a last resort, and the exam expects you to know why.
• ✕Ignoring HCP Terraform because you only use the CLI. It is one of the eight published objectives.

Exam-day tips

• ✓Sixty minutes is short. Do not stall on any single item.
• ✓Read code snippets for the one line that differs between options rather than tracing the whole configuration.
• ✓When two answers both work, prefer the one that keeps state consistent and avoids destroying resources.
• ✓Watch for questions describing team workflows. Those are usually pointing at remote state and locking.
• ✓Answer everything. There is no published penalty for a wrong answer, and a blank certainly scores nothing.
• ✓If a question mentions a version, check whether the behaviour it describes still applies in 1.12.

How to know you’re ready

One good practice score can be luck. What counts is scoring at or above the real pass mark (Not published by HashiCorp) across several full-length sets in a row, with no single domain trailing far behind the others. Kwizza tracks your per-domain readiness automatically as you practice.