Question count and passing score. This course does not quote either.
Total study time
About 26 hours over 4 weeks, at 6–7 h/week. Week 4 is the heaviest at about 7.5 h.
Lab cost
$0. Labs use the random, local, terraform_data, docker and vault providers, the local Docker engine, a Vault dev server, and a free HCP Terraform account. Optional AWS variants are marked.
Start date
Thursday, Oct 15, 2026. Target exam window is Nov 11–13, 2026.
Domain emphasis: Domain 4 (Configuration) gets the most time. 004 adds new emphasis on depends_on, custom conditions, and ephemeral and write-only values, all of which sit in Domain 4. Domains 6 (State) and 8 (HCP Terraform) are next.
Course Overview (1.0 h)
#
Lesson
Time
0.1
Exam facts, 003 → 004 changes, what "provider-agnostic" means for your study
10 min
0.2
Registering for the exam: HashiCorp certification portal, scheduling, ID and system check
10 min
0.3
How to use this course: lesson → lab → knowledge check → cheat sheet
5 min
0.4
Lab 0: Environment setup
35 min
Lab 0 covers these steps:
Install Terraform 1.12.x (official package or tfenv) and confirm with terraform version.
Install VS Code with the HashiCorp Terraform extension.
Install git and create a GitHub account (needed for the VCS lab in Module 8).
Install Docker Engine or Docker Desktop.
Install the Vault binary, to run in dev mode.
Create a free HCP Terraform account. Check the current Free-tier limits when you sign up, because plan contents change.
Create one ~/tf-course/ repo for all labs.
Module 1: Infrastructure as Code (0.75 h) · Objectives 1a–1c
#
Lesson
Obj
Time
1.1
What IaC is: declarative vs. imperative, desired state, idempotency
1a
10 min
1.2
Advantages of IaC: version control, peer review, repeatability, drift reduction, self-documentation
1b
8 min
1.3
Multi-cloud, hybrid, and service-agnostic workflows: one language and workflow, many providers. Terraform vs. config-management tools.
1c
10 min
Lab 1
First config using random_pet and local_file: init, plan, apply, destroy, plus a look at the state file
1a–1c
15 min
KC 1
10 questions
7 min
CS 1
Cheat sheet
Exam traps covered:
Terraform is declarative and does not run steps in file order.
IaC does not remove the need for state.
"Multi-cloud" means multiple providers in one workflow. It does not mean one resource type that works across all clouds.
How Terraform uses providers: plugin architecture, Terraform core vs. provider, the registry, provider source addresses, tiers
2b
15 min
2.2
The terraform block covers three things. It covers required_version vs. required_providers. It covers all version-constraint operators (=, !=, >, >=, <, <=, ~>). It shows what ~> 1.2 vs. ~> 1.2.0 permits.
2a
20 min
2.3
.terraform.lock.hcl covers what the lock file records, when it changes, whether to commit it, and terraform init -upgrade. It also covers terraform providers lock.
2a
15 min
2.4
Multiple providers in one config: provider configurations, alias, the provider meta-argument, preview of passing providers to modules
2c
20 min
2.5
State introduction: why state exists (mapping, metadata, performance), what is stored, and that it is plaintext
2d
15 min
Lab 2
Pin random and local providers, inspect the lock file, loosen a constraint, run init -upgrade, diff the lock file. Configure two aliased provider instances and select one per resource. Optional AWS variant: two regions.
The workflow: Write → Plan → Apply, and the team workflow with PR-based review
3a
10 min
3.2
terraform init downloads providers and modules, initializes the backend, and creates .terraform/. It is safe to re-run.
3b
12 min
3.3
terraform validate checks syntax and internal consistency. It requires init. It does not contact real infrastructure or verify that remote resources exist.
terraform apply: interactive vs. -auto-approve, applying a saved plan without a re-prompt, partial failure behavior
3e
10 min
3.6
terraform destroy vs. apply -destroy vs. plan -destroy, and removing a single resource
3f
8 min
3.7
terraform fmt with -check, -recursive and -diff, plus style conventions. Ends with a matrix of which command touches config, state, or real infrastructure.
3g
12 min
Lab 3
Run the full lifecycle on a Docker nginx container. Save a plan with -out and apply it. Force a replacement with -replace. Use -target and read its warning. Run fmt -check in a CI-style step.
resource vs. data blocks: lifecycle ownership, when data sources are read, terraform_data vs. null_resource
4a
20 min
4.2
Resource addresses and attribute references, including cross-resource references and the implicit dependencies they create
4b
20 min
4.3
Input variables: types, defaults, nullable, sensitive, ephemeral. Full precedence order (env TF_VAR_ → terraform.tfvars → *.auto.tfvars → -var/-var-file). Outputs, including sensitive outputs.
4c
30 min
4.4
Complex types: list, set, map, object, tuple, optional() attributes, and automatic type conversion
4d
25 min
4.5
Expressions: for expressions, splat, conditionals, dynamic blocks, locals. Common functions: lookup, merge, flatten, toset, try, can, coalesce, format, join/split, cidrsubnet, file/templatefile. Using terraform console.
4e
45 min
4.6
Dependencies and meta-arguments. Covers implicit vs. explicit depends_on (004 emphasis), count vs. for_each (and why for_each keys matter), and provider. Covers lifecycle: create_before_destroy, prevent_destroy, ignore_changes, replace_triggered_by.
4f
40 min
4.7
Custom conditions (004 emphasis): validation blocks, precondition/postcondition, check blocks with scoped data sources. Covers when each runs and whether it blocks the apply.
4g
30 min
4.8
Sensitive data (004 emphasis). Covers what sensitive = true does and does not do, that state still holds values in plaintext, ephemeral variables, outputs, and resources, and write-only arguments. Also covers the Vault provider and HCP Terraform dynamic provider credentials.
4h
35 min
Lab 4A
Variables, types and functions: drive local_file contents from objects and maps, test every precedence level, explore expressions in terraform console
4c–4e
25 min
Lab 4B
count → for_each conversion, a forced depends_on, each lifecycle argument demonstrated
4b, 4f
25 min
Lab 4C
Break and fix a variable validation, a precondition, a postcondition, and a check block
4g
20 min
Lab 4D
Vault dev server: write a secret, read it through an ephemeral resource, set a write-only argument, inspect state to confirm the value is absent
4h
25 min
KC 4 / CS 4
20 min
Accuracy flag for Lab 4D. Ephemeral resources and write-only arguments depend on the provider version. Ephemeral resources arrived in Terraform 1.10, and write-only arguments in 1.11. The lab checks the installed provider's schema directly rather than assuming a minimum version.
Module 5: Modules (2.0 h) · Objectives 5a–5d
#
Lesson
Obj
Time
5.1
Module sources: local paths, public and private registry, GitHub, generic Git (with ref), Bitbucket, HTTP, S3, GCS. Shows which sources accept the version argument (registry only).
5a
20 min
5.2
Variable scope: root vs. child modules, why child modules can't see parent variables, and passing data through inputs and outputs
5b
15 min
5.3
Using modules: the module block, count/for_each on modules, providers map and aliases, depends_on on modules
5c
20 min
5.4
Module versions: constraints, Git refs and tags, init -upgrade, why modules aren't recorded in the lock file
5d
15 min
Lab 5
Extract Lab 4 code into a local child module. Call it twice with for_each. Publish it to a GitHub repo and consume it by Git tag.
Local backend: terraform.tfstate, .backup, state contents and sensitivity, why manual edits are discouraged
6a, 2d
15 min
6.2
State locking: which operations lock, backends that support locking, -lock-timeout, terraform force-unlock
6b
15 min
6.3
The backend block. Covers remote backends (s3 with use_lockfile, azurerm, gcs, pg, consul, http) and partial configuration with -backend-config. Covers init -migrate-state vs. -reconfigure, and the terraform_remote_state data source.
6c
25 min
6.4
Drift: refresh-only plan and apply, ignore_changes, -replace. Covers terraform state mv/rm vs. moved/removed blocks.
6d
25 min
6.5
CLI workspaces: terraform workspace new/select/list, terraform.workspace, state file layout, when not to use them. This sets up the comparison in 8.2.
6c/8c
15 min
Lab 6A
Run PostgreSQL in Docker and use it as the pg backend. Migrate local state into it with partial config. Run two terminals to show a lock conflict.
6a–6c
30 min
Lab 6B
Change a managed resource outside Terraform. Detect the drift with -refresh-only. Rename resources with a moved block. Stop managing one with a removed block.
Importing existing infrastructure: terraform import CLI vs. import blocks (plannable, for_each support), plan -generate-config-out
7a
20 min
7.2
Inspecting state: state list, state show, show, show -json, output (-json, -raw), graph, providers
7b
15 min
7.3
Verbose logging: TF_LOG levels (TRACE, DEBUG, INFO, WARN, ERROR), TF_LOG_PATH, TF_LOG_CORE vs. TF_LOG_PROVIDER, and when logs help
7c
10 min
Lab 7
Create Docker networks by hand. Import them with the CLI, with an import block and -generate-config-out, and with a for_each import. Capture a TRACE log to a file.
Editions: Community edition (CLI) vs. HCP Terraform (SaaS) vs. Terraform Enterprise (self-hosted). Covers the org → project → workspace hierarchy.
8a–8c
15 min
8.2
Workspaces and projects: HCP workspaces vs. CLI workspaces (a classic exam trap), project-level permissions, workspace tags
8c
20 min
8.3
Integration: terraform login, the cloud block vs. a backend block, migrating state. Covers CLI-driven, VCS-driven and API-driven runs. Covers remote, local and agent execution modes.
8a, 8d
25 min
8.4
Variables in HCP Terraform: workspace variables (Terraform vs. environment variables, sensitive), variable sets, precedence
8a, 8d
15 min
8.5
Collaboration and governance: teams and permissions, private registry, Sentinel and OPA policies, run tasks, health assessments and drift detection, run triggers.
8b
25 min
Lab 8A
Migrate a config into HCP Terraform with the cloud block. Run a CLI-driven remote plan and apply. Create a variable set.
8a, 8c, 8d
30 min
Lab 8B
Connect a GitHub repo for a VCS-driven workspace. Open a PR to get a speculative plan. Merge it to trigger an apply. Policy and health-assessment features are a read-only walkthrough if your tier doesn't include them.
8a–8d
30 min
KC 8 / CS 8
20 min
Accuracy flag for 8.5. Which features are in which HCP Terraform tier changes over time. Feature availability is marked "check your plan," and exam questions test concepts, not plan names.
Assessment and Wrap-Up
Section
Contents
Time
Practice Exam 1
60 questions, 60-minute timer, mixed single-answer, multi-select and true/false
1 h + 0.75 h review
Practice Exam 2
60 new questions with the same blueprint
1 h + 0.75 h review
Readiness Assessment
Score bands, a remediation map from missed objective IDs to lessons and labs, and a go/no-go rule: schedule only after scoring at least 85% on both exams
Used after each exam
Exam-Day Guide
Proctoring and room requirements, pacing, how to read questions, flag-and-return
15 min
Glossary and Command Reference
Every term and CLI command in the course, with objective tags
Reference
Practice exam blueprint. This weighting is my own design for practice. HashiCorp does not publish domain percentages.
Domain
1
2
3
4
5
6
7
8
Total
Questions per exam
4
7
7
14
6
9
5
8
60
Each exam is about 70% single-answer, 20% multi-select, and 10% true/false.