kb/core

Terraform basics: init, plan, apply

The Terraform basics cycle of init, plan, and apply: HCL describes the desired state, `terraform init` installs the providers, `plan` shows the diff against state, and `apply` makes the changes. This is the foundation. Without the cycle you cannot work with any provider.

tf-interpolation${...}: substituting values into strings
${expression} inside a string substitutes the value of that expression. ${var.env}, ${aws_s3_bucket.demo.id}. When the expression is the entire value of an argument (no surrounding string), modern HCL lets you write it without ${...}.
tf-lockfile.terraform.lock.hcl: pinning provider versions
The lockfile pins the exact provider versions and their hashes, so you and CI always run the same build. It is created on terraform init and updated through init -upgrade. Commit it to git.
tf-init-backendsBackends in Terraform: where state lives
A backend is where the state file is stored. The default is local, next to your HCL. Remote backends (S3, GCS, Terraform Cloud, http) give you shared access and locking. This course uses only local; remote is covered as an overview.
tf-conditional-expressionConditionals and null safety: ?:, try, can, coalesce
The ternary operator a ? b : c is a plain if/else. try(expr, fallback) evaluates an expression or substitutes a backup. can(expr) returns true/false. coalesce(...) returns the first non-null value. They all keep your config from crashing.
hcl-typesData types in HCL: string, number, list, map, object
HCL supports primitives (string, number, bool) and complex types: list, set, map, tuple, object. This article covers the syntax of each one and the difference between the look-alikes (list vs tuple, map vs object).
tf-functions-collectionHCL collection functions: length, lookup, merge, concat, flatten
HCL has functions for list/set/map: length (size), lookup (by key with a default), merge (combine maps), concat (join lists), flatten (unnest), keys/values. The basic toolkit for transformations.
tf-functions-stringHCL string functions: format, join, replace, lower, and more
HCL string functions: format (like printf), join/split (lists to strings and back), replace (regex or text), lower/upper/title (case), trimspace and the trim* family. All pure, no side effects, handy in locals.
hcl-syntaxHCL: the language you write Terraform in
HCL (HashiCorp Configuration Language) is the language you use to describe the desired state of your infrastructure. It looks like JSON, but it is easier to read: you can write comments, variables, and loops.
tf-init-modulesHow terraform init pulls in modules
When your HCL has a module block, terraform init downloads the module source (from a registry, git, or a local folder) into .terraform/modules/. This course does not write modules; this article is an overview of how the mechanics work.
tf-referencesReferences in HCL: how to read aws_s3_bucket.demo.bucket
Any value in HCL is reachable through an address: var.x, local.x, aws_s3_bucket.demo.arn, module.net.vpc_id, data.aws_region.current.name. Knowing this syntax is half of your productivity in Terraform.
tf-applyterraform apply: apply a plan to a real cloud
apply takes the result of plan and actually calls the cloud API: it creates, changes, and deletes resources. After apply, the state is updated. This is the command that turns money into infrastructure.
tf-cli-configTerraform CLI configuration: terraformrc, env vars, TF_LOG
You configure the Terraform CLI through the ~/.terraformrc file and the TF_* environment variables. This is where the plugin cache lives, along with TF_LOG for debugging, TF_VAR_* for variables, and TF_CLI_ARGS for global flags.
tf-destroyterraform destroy: tear down everything you created
destroy removes every resource described in HCL and recorded in state. It is apply with a minus sign in front of everything. It is essential for learning tasks and ephemeral environments. In production it is a last resort.
tf-initterraform init: the first command in any project
terraform init downloads the provider plugins (AWS, GCP, and so on), creates a lockfile that pins their versions, and prepares the working directory. Without it, neither plan nor apply will run.
tf-planterraform plan: see what Terraform is about to do
plan is a dry run: Terraform reads your HCL, reads the state, and shows the diff between them. It changes nothing in the cloud. This is your main tool for not breaking prod by mistake.
tf-workspaceterraform workspace: several states in one directory
A workspace is a named slot for a separate state file in one directory of HCL. It is useful for very similar environments with a minimal difference between them. This course does not use it; separate folders are usually the better choice.
tf-version-constraintsVersion constraints in Terraform: required_version and providers
required_version pins which versions of terraform may run this code. required_providers.version does the same for providers. The pessimistic operator ~> 5.60 is the standard: it allows minor updates and blocks major ones.

kb/core

Terraform basics: init, plan, apply

tf-interpolation${...}: substituting values into strings
${expression} inside a string substitutes the value of that expression. ${var.env}, ${aws_s3_bucket.demo.id}. When the expression is the entire value of an argument (no surrounding string), modern HCL lets you write it without ${...}.
tf-lockfile.terraform.lock.hcl: pinning provider versions
The lockfile pins the exact provider versions and their hashes, so you and CI always run the same build. It is created on terraform init and updated through init -upgrade. Commit it to git.
tf-init-backendsBackends in Terraform: where state lives
A backend is where the state file is stored. The default is local, next to your HCL. Remote backends (S3, GCS, Terraform Cloud, http) give you shared access and locking. This course uses only local; remote is covered as an overview.
tf-conditional-expressionConditionals and null safety: ?:, try, can, coalesce
The ternary operator a ? b : c is a plain if/else. try(expr, fallback) evaluates an expression or substitutes a backup. can(expr) returns true/false. coalesce(...) returns the first non-null value. They all keep your config from crashing.
hcl-typesData types in HCL: string, number, list, map, object
HCL supports primitives (string, number, bool) and complex types: list, set, map, tuple, object. This article covers the syntax of each one and the difference between the look-alikes (list vs tuple, map vs object).
tf-functions-collectionHCL collection functions: length, lookup, merge, concat, flatten
HCL has functions for list/set/map: length (size), lookup (by key with a default), merge (combine maps), concat (join lists), flatten (unnest), keys/values. The basic toolkit for transformations.
tf-functions-stringHCL string functions: format, join, replace, lower, and more
HCL string functions: format (like printf), join/split (lists to strings and back), replace (regex or text), lower/upper/title (case), trimspace and the trim* family. All pure, no side effects, handy in locals.
hcl-syntaxHCL: the language you write Terraform in
HCL (HashiCorp Configuration Language) is the language you use to describe the desired state of your infrastructure. It looks like JSON, but it is easier to read: you can write comments, variables, and loops.
tf-init-modulesHow terraform init pulls in modules
When your HCL has a module block, terraform init downloads the module source (from a registry, git, or a local folder) into .terraform/modules/. This course does not write modules; this article is an overview of how the mechanics work.
tf-referencesReferences in HCL: how to read aws_s3_bucket.demo.bucket
Any value in HCL is reachable through an address: var.x, local.x, aws_s3_bucket.demo.arn, module.net.vpc_id, data.aws_region.current.name. Knowing this syntax is half of your productivity in Terraform.
tf-applyterraform apply: apply a plan to a real cloud
apply takes the result of plan and actually calls the cloud API: it creates, changes, and deletes resources. After apply, the state is updated. This is the command that turns money into infrastructure.
tf-cli-configTerraform CLI configuration: terraformrc, env vars, TF_LOG
You configure the Terraform CLI through the ~/.terraformrc file and the TF_* environment variables. This is where the plugin cache lives, along with TF_LOG for debugging, TF_VAR_* for variables, and TF_CLI_ARGS for global flags.
tf-destroyterraform destroy: tear down everything you created
destroy removes every resource described in HCL and recorded in state. It is apply with a minus sign in front of everything. It is essential for learning tasks and ephemeral environments. In production it is a last resort.
tf-initterraform init: the first command in any project
terraform init downloads the provider plugins (AWS, GCP, and so on), creates a lockfile that pins their versions, and prepares the working directory. Without it, neither plan nor apply will run.
tf-planterraform plan: see what Terraform is about to do
plan is a dry run: Terraform reads your HCL, reads the state, and shows the diff between them. It changes nothing in the cloud. This is your main tool for not breaking prod by mistake.
tf-workspaceterraform workspace: several states in one directory
A workspace is a named slot for a separate state file in one directory of HCL. It is useful for very similar environments with a minimal difference between them. This course does not use it; separate folders are usually the better choice.
tf-version-constraintsVersion constraints in Terraform: required_version and providers
required_version pins which versions of terraform may run this code. required_providers.version does the same for providers. The pessimistic operator ~> 5.60 is the standard: it allows minor updates and blocks major ones.