Terraform-Associate-003 HashiCorp Certified: Terraform Associate (003) (HCTA0-003) Question and Answers

This command will initialize the new backend and prompt you to migrate the existing state file to the new location4. The other commands are not relevant for this task.

You can develop a custom provider to manage its resources using Terraform, as Terraform is an extensible tool that allows you to write your own plugins in Go language. You can also publish your custom provider to the Terraform Registry or use it privately.

Destroy Command Options: Theterraform destroycommand is the correct method to destroy resources, and it requires either manual approval or the -auto-approve flag.

Unsupported Triggering Method: The--destroy flagis not a recognized option for plan or apply commands, making B the correct answer as it isnota valid way to initiate resource destruction.

For more on destroying resources, refer to theterraform destroy command documentationin the Terraform CLI reference.

In Terraform, theresource nameis the second argument in the resource block.

The structure of a resource block is:

resource "provider_resource_type" "resource_name" {

# Configuration settings

}

Here, the provider type isgoogle_compute_instance, and theresource nameis "main".

The name "test" is simply the value assigned to the name attribute, which is unrelated to the Terraform resource name.

"google" and "compute_instance" are part of the provider and resource type, not the resource name.

Official Terraform Documentation Reference:

Terraform Resource Documentation

terraform validatechecks for syntax errors and internal consistencyin the configuration files.

However, itdoes notcheck if resource configurations are valid with the provider.

Official Terraform Documentation Reference:

terraform validate - HashiCorp Documentation

The key principle of infrastructure as code that is not listed among the options is golden images. Golden images are pre-configured, ready-to-use virtual machine images that contain a specific set of software and configuration. They are often used to create multiple identical instances of the same environment, such as for testing or production. However, golden images are not a principle of infrastructure as code, but rather a technique that can be used with or without infrastructure as code. The other options are all key principles of infrastructure as code, as explained below:

Self-describing infrastructure: This means that the infrastructure is defined in code that describes its desired state, rather than in scripts that describe the steps to create it. This makes the infrastructure easier to understand, maintain, and reproduce.

Idempotence: This means that applying the same infrastructure code multiple times will always result in the same state, regardless of the initial state. This makes the infrastructure consistent and predictable, and avoids errors or conflicts caused by repeated actions.

Versioned infrastructure: This means that the infrastructure code is stored in a version control system, such as Git, that tracks the changes and history of the code. This makes the infrastructure code reusable, auditable, and collaborative, and enables practices such as branching, merging, and rollback. References = [Introduction to Infrastructure as Code with Terraform], [Infrastructure as Code in a Private or Public Cloud]

Terraform can manage resource dependencies implicitly or explicitly. Implicit dependencies are created when a resource references another resource or data source in its arguments. Terraform can infer the dependency from the reference and create or destroy the resources in the correct order. Explicit dependencies are created when you use the depends_on argument to specify that a resource depends on another resource or module. This is useful when Terraform cannot infer the dependency from the configuration or when you need to create a dependency for some reason outside of Terraform’s scope. References = : Create resource dependencies : Terraform Resource Dependencies Explained

Sentinel policies are checked after the plan stage of a Terraform run, but before it can be confirmed or the terraform apply is executed3. This allows you to enforce rules on your infrastructure before it is created or modified.

Speculative Plans: Terraform Cloud’sspeculative planfeature runs automatically when changes are detected in a linked VCS repository, enabling users to review potential infrastructure changes without committing them.

Automatic Integration: This feature automates the planning process by triggering when changes are committed, aiding teams in previewing infrastructure changes seamlessly.

For further understanding, see theTerraform Cloud VCS Integrationdocumentation.

In Terraform, the correct way to reference a resource attribute is:

pgsql

CopyEdit

resource_type.resource_name.attribute

For example, if the resource block is:

hcl

CopyEdit

resource "kubernetes_namespace" "example" {

metadata {

name = "my-namespace"

}

}

To reference the name attribute, use:

pgsql

CopyEdit

kubernetes_namespace.example.name

Explanation of incorrect answers:

A (resource.kubernetes_namespace.example.name)– Incorrect. Terraform does not use the resource. prefix when referencing resources.

C (data.kubernetes.namespace.name)– Incorrect. This syntax is used fordata sources, not resources.

D (kubernetes_namespace.test.name)– Incorrect. The resource name is "example", not "test".

Official Terraform Documentation Reference:

Terraform Resource References

terraform validate does not confirm that your infrastructure matches the Terraform state file. It only checks whether the configuration files in a directory are syntactically valid and internally consistent3. To confirm that your infrastructure matches the Terraform state file, you need to use terraform plan or terraform apply with the -refresh-only option.

C (✅Correct)– If changes require aresource replacement(e.g., changing an immutable attribute like instance type), Terraform willdestroy and recreate the resource.

E (✅Correct)– Terraform only appliesthe configuration in the current directory or workspace.

From the Terraform Output Docs – Sensitive:

"Marking an output as sensitive prevents Terraform from showing its value in the CLI output,but it will still bestored in the state file."

Thus, it protects display, not storage.

A (✅Correct)– Providers allow Terraform tointeract with APIsof cloud/on-premises services.

B (✅Correct)– Some Terraform providers can provisionon-premises infrastructure, such as VMware, OpenStack, etc.

C (❌Incorrect)– This describesTerraform Workspaces, not providers.

D (✅Correct)– Terraform providers allow provisioning ofpublic cloud resources(AWS, Azure, GCP, etc.).

E (❌Incorrect)– Enforcing security and compliance policies isnot a direct provider function, but it can be done using Sentinel or other policy-as-code tools.

Official Terraform Documentation Reference:

Terraform Providers

Detailed Explanation:

Rationale for Correct Answer (B):IaC best practice is to manage infrastructure through version-controlled code. Changes should be reviewed and approved (via PRs), ensuring collaboration, traceability, and automation.

Analysis of Incorrect Options:

A, D, E: Making direct/manual changes bypasses IaC practices and causes drift.

C: Running code without PR review skips collaboration and approval.

Key Concept:Infrastructure as Code emphasizes version control + peer review + automation.

Automated Visualization: HCP Terraform providesvisualization toolsthat map infrastructure configurations, helping users manage complex architectures.

Remote State Storage: Terraform Cloud offersremote state management, essential for teams working collaboratively on shared infrastructure, ensuring consistency and avoiding state conflicts.

For more information, consultTerraform Cloud and HCP Terraform featuresin the official documentation.

Detailed Explanation:

Rationale for Correct Answer (A):Providers only interact with one specific platform or API. Terraform itself can work with multiple providers in one configuration, but a single provider is not responsible for provisioning across multiple cloud providers.

Analysis of Incorrect Options:

B. Managing actions to take based on resource differences: This is a provider responsibility (through the resource schema).

C. Managing resources and data sources based on an API: Correct — providers define resources/data sources and map them to API calls.

D. Understanding API interactions with a hosted service: Correct — providers encapsulate API logic to allow Terraform to manage resources.

Key Concept:Providers are API adapters. They handle CRUD operations for resources in their own ecosystem but do not span across multiple cloud platforms.

This will trigger a run in the Terraform Cloud workspace, which will perform a plan and apply operation on the infrastructure defined by the Terraform configuration files in the VCS repository.

Detailed Explanation:

Rationale for Correct Answer (True):When terraform apply completes successfully, Terraform prints output values. Outputs from both root and child modules are displayed, but child module outputs must be explicitly exposed through the root module outputs to be visible at the CLI.

Analysis of Incorrect Option:

False: Incorrect, because Terraform does display output values, but only if they are exposed from child modules to the root module.

Key Concept:Outputs help you extract important information (e.g., IP addresses, resource IDs) from your configuration.

The command that makes your code more human readable is terraform fmt. This command is used to rewrite Terraform configuration files to a canonical format and style, following the Terraform language style conventions and other minor adjustments for readability. The command is optional, opinionated, and has no customization options, but it is recommended to ensure consistency of style across different Terraform codebases. Consistency can help your team understand the code more quickly and easily, making the use of terraform fmt very important. You can run this command on your configuration files before committing them to source control or as part of your CI/CD pipeline. References = : Command: fmt : Using Terraform fmt Command to Format Your Terraform Code

This is not true, as modules can be either public or private, depending on your needs and preferences. You can use the Terraform Registry to publish and consume public modules, or use Terraform Cloud or Terraform Enterprise to host and manage private modules.

The provider for theaws_vpcresource isaws, as the resource type begins withaws_, which denotes that it is managed by the AWS provider.

This is the Terraform style convention for indenting a nesting level compared to the one above it. The other options are not consistent with the Terraform style guide.

You cannot access state stored with the local backend by using the terraform_remote_state data source. The terraform_remote_state data source is used to retrieve the root module output values from some other Terraform configuration using the latest state snapshot from the remote backend. It requires a backend that supports remote state storage, such as S3, Consul, AzureRM, or GCS. The local backend stores the state file locally on the filesystem, which terraform_remote_state cannot access.

The terraform state list command lists all resources that are managed by Terraform in the current state file1. The terraform state show command shows the attributes of a single resource in the state file2. By using these two commands, you can compare the VM IDs in your list with the ones in the state file and identify which one is managed by Terraform.

A module cannot always refer to all variables declared in its parent module, as it needs to explicitly declare input variables and assign values to them from the parent module’s arguments. A module cannot access the parent module’s variables directly, unless they are passed as input arguments.

Infrastructure as code (IaC) is a way of managing and provisioning cloud infrastructure using programming techniques instead of manual processes1. IaC has many advantages over using a graphical user interface (GUI) for provisioning infrastructure, such as:

•Versioning: IaC allows you to store your infrastructure configuration in a version control system, such as Git, and track changes over time. This enables you to roll back to previous versions, compare differences, and collaborate with other developers2.

•Reusability: IaC allows you to create reusable modules and templates that can be applied to different environments, such as development, testing, and production. This reduces duplication, improves consistency, and speeds up deployment3.

•Sharing: IaC allows you to share your infrastructure configuration with other developers, teams, or organizations, and leverage existing code from open source repositories or registries. This fosters best practices, innovation, and standardization4.

•Risk reduction: IaC reduces the risk of human error, configuration drift, and security breaches that can occur when provisioning infrastructure manually or using a GUI. IaC also enables you to perform automated testing, validation, and compliance checks on your infrastructure before deploying it5.

References =

•1: What is Infrastructure as Code? Explained for Beginners - freeCodeCamp.org

•2: The benefits of Infrastructure as Code - Microsoft Community Hub

•3: Infrastructure as Code : Best Practices, Benefits & Examples - Spacelift

•4: 5 Benefits of Infrastructure as Code (IaC) for Modern Businesses in the Cloud

•5: The 7 Biggest Benefits of Infrastructure as Code - DuploCloud

This is the kind of configuration block that will create an infrastructure object with settings specified within the block. The other options are not used for creating infrastructure objects, but for configuring providers, accessing state data, or querying data sources.

 The concept of infrastructure as code (IaC) is to define and manage infrastructure using code, rather than manual processes or GUI tools. A script that contains a series of public cloud CLI commands is an example of IaC, because it uses code to provision resources into a public cloud. The other options are not examples of IaC, because they involve manual or interactive actions, such as running curl commands, sending REST requests, or entering commands into a console. References = [Introduction to Infrastructure as Code with Terraform] and [Infrastructure as Code]

Terraformdownloads modules and providersinto the .terraform directory inside the working directory.

A (/tmp/)– Incorrect; modules arenot stored temporarily.

C (In memory)– Incorrect; modules persist on disk.

D (Not cached)– Incorrect; Terraformdoes cache modulesfor efficiency.

Official Terraform Documentation Reference:

Terraform Module Caching

The terraform validate command is used to check that your Terraform configuration files are syntactically valid and internally consistent. It is a useful command for ensuring your Terraform code is error-free before applying any changes to your infrastructure.

You can install third party plugins using terraform init, as long as you specify the plugin directory in your configuration or as a command-line argument. You can also use the terraform providers mirror command to create a local mirror of providers from any source.

This command will check if all code in a Terraform configuration that references multiple modules is properly formatted without making changes, and will return a non-zero exit code if any files need formatting. The other commands will either make changes, list the files that need formatting, or not check the modules.

Environment variables and connection configurations outside of Terraform are secure options for storing secrets for connecting to a Terraform remote backend. Environment variables can be used to set values for input variables that contain secrets, such as backend access keys or tokens. Terraform will read environment variables that start with TF_VAR_ and match the name of an input variable. For example, if you have an input variable called backend_token, you can set its value with the environment variable TF_VAR_backend_token1. Connection configurations outside of Terraform are files or scripts that provide credentials or other information for Terraform to connect to a remote backend. For example, you can use a credentials file for the S3 backend2, or a shell script for the HTTP backend3. These files or scripts are not part of the Terraform configuration and can be stored securely in a separate location. The other options are not secure for storing secrets. A variable file is a file that contains values for input variables. Variable files are usually stored in the same directory as the Terraform configuration or in a version control system. This exposes the secrets to anyone who can access the files or the repository. You should not store secrets in variable files1. Inside the backend block within the Terraform configuration is where you specify the type and settings of the remote backend. The backend block is part of the Terraform configuration and is usually stored in a version control system. This exposes the secrets to anyone who can access the configuration or the repository. You should not store secrets in the backend block4. References = [Terraform Input Variables]1, [Backend Type: s3]2, [Backend Type: http]3, [Backend Configuration]4

The terraform plan command is used to create an execution plan. It allows you to see what actions Terraform will take to reach the desired state defined in your configuration files. It evaluates the current state and configuration, showing a detailed outline of the resources that will be created, updated, or destroyed. This is a critical step in the development process as it helps you verify that the changes you are about to apply will perform as expected, without actually modifying any state or infrastructure.

Changing the Terraform backend after performing the initial terraform apply is technically possible but strongly discouraged. This is because changing backends can lead to complexities in state management, requiring manual intervention such as state migration to ensure consistency. Terraform's documentation and best practices advise planning the backend configuration carefully before applying Terraform configurations to avoid such changes.References = This guidance is consistent with Terraform's official documentation, which recommends careful consideration and planning of backend configurations to avoid the need for changes.

The purpose of state in Terraform is to keep track of the real-world resources managed by Terraform, mapping them to the configuration. The state file contains metadata about these resources, such as resource IDs and other important attributes, which Terraform uses to plan and manage infrastructure changes. The state enables Terraform to know what resources are managed by which configurations and helps in maintaining the desired state of the infrastructure.References = This role of state in Terraform is outlined in Terraform's official documentation, emphasizing its function in mapping configuration to real-world resources and storing vital metadata .

This is where Terraform cannot load a provider from, as it requires a compiled binary file that implements the provider protocol. You can load a provider from a plugins directory, a provider plugin cache, or the official HashiCorp distribution on releases.hashicorp.com.

Module version is optional to reference a module on the Terraform Module Registry. If you omit the version constraint, Terraform will automatically use the latest available version of the module

Terraform can call modules from various sources including the public Terraform Registry, private registries, local file paths, or version control systems like GitHub.

The execution plan for terraform apply will not be the same as the one you ran locally with terraform plan, if your teammate manually modified the infrastructure component you are working on. This is because Terraform will refresh the state file before applying any changes, and will detect any differences between the state and the real resources.

 This is the command that can add existing resources into Terraform state, by matching them with the corresponding configuration blocks in your files. 

The Terraform Cloud private registry provides the ability to restrict modules to members of Terraform Cloud or Enterprise organizations. This allows you to share modules within your organization without exposing them to the public. The private registry also supports importing modules from your private VCS repositories. The public Terraform Module Registry, on the other hand, publishes modules from public Git repositories and makes them available to any user of Terraform. References = : Private Registry - Terraform Cloud : Terraform Registry - Provider Documentation

To output returned values from a child module in the Terraform CLI output, you need to declare the output in both the child module and the root module. The child module output will return the value to the root module, and the root module output will display the value in the CLI. References = [Terraform Outputs]

This is the recommended way to use a remote backend that needs authentication, as it allows you to provide the credentials via environment variables, command-line arguments, or interactive prompts, without storing them in the Terraform configuration files.

This is how the Terraform Cloud integration differs from other state backends such as S3, Consul, etc., as it allows you to perform remote operations on Terraform Cloud’s servers instead of your local machine. The other options are either incorrect or irrelevant.

Purpose of Refresh-Only Mode: Running Terraform inrefresh-only modeupdates Terraform's state file with the current state of resources in your infrastructure without making changes to the resources themselves.

Context of Terraform Import: When usingterraform import, you’re adding existing resources to the state file, and running Terraform in refresh-only mode before this operation can ensure that any initial configuration syncs precisely with the actual state.

For more on refresh-only mode in relation to terraform import, refer to Terraform's import documentation.

When the state file is locked, operations that modify or depend on the state (liketerraform apply,terraform destroy, andterraform state list) are blocked.terraform fmtonly formats the configuration files and does not interact with the state, so it is allowed.

Terraform Cloud provides features like remote state storage and a web-based user interface for managing your Terraform runs. While it offers robust infrastructure as code capabilities, automatic backups of configuration and state are not directly provided by Terraform Cloud; instead, the state is stored remotely and secured.

Runningterraform destroywill show all resources that will be deleted before prompting for approval. You can also runterraform plan -destroyto simulate the destruction without actually applying it, which is useful for reviewing the planned changes.

To see all the resources that Terraform will delete, you can use either of these two commands:

terraform destroy will show the plan of destruction and ask for your confirmation before proceeding. You can cancel the command if you do not want to destroy the resources.

terraform plan -destroy will show the plan of destruction without asking for confirmation. You can use this command to review the changes before running terraform destroy. References = : Destroy Infrastructure : Plan Command: Options

The terraform state file is locked when a Terraform operation that could write state is in progress. This prevents concurrent state operations that could corrupt the state. The forbidden actions when the state file is locked are those that could write state, such as terraform apply, terraform destroy, terraform refresh, terraform taint, terraform untaint, terraform import, and terraform state *. The terraform validate command is also forbidden, because it requires an initialized working directory with the state file. The allowed actions when the state file is locked are those that only read state, such as terraform plan, terraform show, terraform output, and terraform console. References = [State Locking] and [Command: validate]

State locking prevents other users from modifying the state file while a Terraform operation is in progress. This prevents conflicts and data loss1.

This is how Terraform chooses which version of the provider to use, when you add a new resource to an existing Terraform configuration, but do not update the version constraint in the configuration. The lock file records the exact version of each provider that was installed in your working directory, and ensures that Terraform will always use the same provider versions until you run terraform init -upgrade to update them.

These are examples of infrastructure as code (IaC), which is a practice of managing and provisioning infrastructure through machine-readable definition files, rather than physical hardware configuration or interactive configuration tools.

Before using a remote backend in Terraform, it is mandatory to run terraform init. This command initializes a Terraform working directory, which includes configuring the backend. If a remote backend is specified, terraform init will set up the working directory to use it, including copying any existing state to the remote backend if necessary.References = This principle is a fundamental part of workingwith Terraform and its backends, as outlined in general Terraform documentation and best practices. The specific HashiCorp Terraform Associate (003) study materials in the provided files did not include direct references to this information.

According to theofficial Terraform documentationhere:

🔗Terraform Docs – Declaring an Output Value

In Terraform 0.12 and later, you can directly use expressions in outputs without interpolation syntax:

output "instance_ip_addr" {

value = aws_instance.server.private_ip

}

This aligns perfectly withOption A:

output "instance_ip_addr" {

value = aws_instance.main.private_ip

}

Why not the others?

❌Option B: Incorrect syntax. "${main.private_ip}" is referencing main as if it were a global variable or resource, but it isn’t defined. Plus, the output name is oddly written as aws_instance.instance_ip_addr, which is not a valid identifier format.

❌Option C: Although valid in older versions (<= 0.11), interpolation with "${}" isdeprecatedin Terraform 0.12+. The docs clearly advise using direct expressions instead.

❌Option D: Terraform hasno return statement; this is syntactically invalid in Terraform’s HCL.

This is what determines how Terraform creates, updates, or deletes resources, as it is responsible for understanding API interactions with some service and exposing resources and data sources based on that API.

Detailed Explanation:

Rationale for Correct Answer:

init: Initializes the working directory, downloads providers and modules.

apply: Executes the planned changes and provisions infrastructure.

Both steps are mandatory for provisioning new infrastructure.

Analysis of Incorrect Options (Distractors):

A. import: Used for existing resources, not provisioning new ones.

B. plan: Optional but recommended; not strictly required.

C. validate: Checks syntax only; does not provision resources.

Key Concept:The Terraform workflow requires initialization and application to create infrastructure.

Detailed Explanation:

Rationale for Correct Answer (D):The terraform.lock.hcl file is automatically created and maintained by Terraform. It records the specific versions and checksums of providers used in a configuration, ensuring consistent runs across environments and machines.

Analysis of Incorrect Options:

A. There is no such file: Incorrect — the lock file exists and is crucial for dependency management.

B. Storing references to workspaces: Incorrect — workspaces are tracked in the state file, not the lock file.

C. Preventing Terraform runs: Incorrect — the lock file does not block runs; it enforces consistent provider versions.

Key Concept:The terraform.lock.hcl file enforces provider version consistency, which is critical for reproducible and reliable Terraform deployments.

Child modulesdo not automatically inheritvariables from the parent module.

To pass values from theparent moduleto thechild module, you mustexplicitly define input variablesin the child module and pass them in the parent module.

Example:

hcl

CopyEdit

module "example" {

source = "./child_module"

var1 = "value"

}

Official Terraform Documentation Reference:

Passing Variables to Modules

A provider configuration block is not required in every Terraform configuration. A provider configuration block can be omitted if its contents would otherwise be empty. Terraform assumes an empty default configuration for any provider that is not explicitly configured. However, some providers may require some configuration arguments (such as endpoint URLs or cloud regions) before they can be used. A provider’s documentation should list which configuration arguments it expects. For providers distributed on the Terraform Registry, versioned documentation is available on each provider’s page, via the “Documentation” link in the provider’s header1. References = [Provider Configuration]1

The public Terraform Module Registry is free to use, as it is a public service that hosts thousands of self-contained packages called modules that are used to provision infrastructure. You can browse, use, and publish modules to the registry without any cost.

In Terraform, the backend configuration, which includes details about where and how state is stored, is specified within the terraform block of your configuration. This block is the correct place to define the backend type and its configuration parameters, such as the location of the state file for a local backend or the bucket details for a remote backend like S3.References = This practice is outlined in Terraform's core documentation, which provides examples and guidelines on how to configure various aspects of Terraform's behavior, including state backends .

Detailed Explanation:

Rationale for Correct Answer:Terraform supports downloading modules from multiple sources, including:

Local file paths

Git repositories (e.g., GitHub)

Terraform Registry

Direct HTTP URLs

Therefore, all listed sources are valid.

Analysis of Incorrect Options (Distractors):

A–D: Each is valid individually, but incomplete alone.

E: Correct because it encompasses all supported sources.

Key Concept:Terraform modules are flexible and source-agnostic, supporting multiple retrieval methods.

This is what will happen if you run terraform destroy without any flags, as it will attempt to delete all the resources that are associated with your current working directory or workspace. You can use the -target flag to specify a particular resource that you want to destroy.

To make Terraform's logging more verbose for troubleshooting purposes, you must configure the TF_LOG environment variable. This variable controls the level of logging and can be set to TRACE, DEBUG, INFO, WARN, or ERROR, with TRACE providing the most verbose output.References = Detailed debugging instructions and the use of environment variables like TF_LOG for increasing verbosity are part of Terraform's standard debugging practices

Detailed Explanation:

Rationale for Correct Answer (C):Credentials should not be hardcoded in Terraform files. Best practice is to configure them via environment variables or secret managers.

Analysis of Incorrect Options:

A: Shared servers introduce risk and drift.

B: Storing secrets in config/version control is insecure.

D: Incorrect since option C is valid.

Key Concept:Separation of credentials from code is a Terraform security best practice.

This is the type of block that is used to construct a collection of nested configuration blocks, by using a for_each argument to iterate over a collection value and generate a nested block for each element. For example, you can use a dynamic block to create multiple ingress rules for a security group resource.

Dynamic blocks in Terraform allow you to define multiple nested blocks within a resource based on the values of a variable. This feature is particularly useful for scenarios where the number of nested blocks is not fixed and can change based on variable input.

The terraform import command is not part of any other Terraform workflow. It must be explicitly invoked by the user with the appropriate arguments, such as the resource address and the ID of the existing infrastructure to import. References = [Importing Infrastructure]

The terraform state rm commandremovesa resource from Terraform’s state file butdoes not destroythe resource in the actual infrastructure. It only removes Terraform’s knowledge of the resource, meaning Terraform will no longer manage it.

If you run terraform state rm on a resource, Terraform will forget that the resource exists.

However, the resource will still exist in the cloud or infrastructure provider.

If you later run terraform apply, Terraform may try torecreatethe resource because it is no longer present in its state file.

Official Terraform Documentation Reference:

terraform state rm - HashiCorp Documentation