CHEATSHEET.md

Cheat sheet

This cheatsheet contains the most commonly used Hashicorp Configuration Language (HCL) syntax, which might come in handy during the exercises.

Resources, data sources and modules

Resources

Resources are the most fundamental objects within HCL; they are the building blocks which you can use to build your infrastructure. Resources have input arguments that you use to configure them, and have output arguments or so-called attributes which you can use to configure other resources and create implicit dependencies. You use them as follows:

resource "cloud_resource" "my_resource" {
    arg1 = x
}

Here you are creating an instance of a cloud_resource type resource (which should be documented in a provider doc), and you give it the name my_resource which you can use to refer to this instance in other parts of your code. Resources are namespaced by their type (note that this is an important difference with other Terraform types!). You can refer to an attribute of a resource (to configure another resource, or to use an output, see below) as follows:

cloud_resource.my_resource.property

This resolves into the attribute attribute of the cloud_resource type that you named my_resource in your code.

Data sources

Note that data sources, like resources, are objects that have attributes that contain the data you are likely interested in. As for resources, you can find the list of attributes of a data source in its Terraform registry entry.

data "cloud_data_source" "my_data_source" {
    arg1 = x
}

Data sources have their own namespace, meaning that you can reference them as follows: data.cloud_data_source.my_data_source.property, which extracts the property attribute of the cloud_data_source that is known to Terraform as my_data_source.

Modules

Modules are directories containing files with .tf prefixes, and can be instantiated in other modules as follows:

module "my_module" {
    source = "./my_module_folder"
    var1 = x
    var2 = y
    var3 = z
}

You can reference modules and the outputs that they contain in a separate module namespace. E.g.: module.my_module.my_module_output refers to the my_module_output output defined in the module.

---

Variables, locals and outputs

Variables

You can define variables as follows:

variable "my_variable" {
    type = list(string)
    default = ["Hello", "world!"]
}

They can be referenced via the var namespace as follows: var.my_variable. Arguments such as type and default are optional. For a complete list of (optional) variable arguments, have a look at the official docs.

Locals

You can define one or more local values in a locals scope:

locals {
    my_first_local = "my_first_local_value"
    my_second_local = 42
}

and they can be referenced as follows: local.my_first_local. Notice that you define them as locals, but the namespace they reside in is called local (without the extra 's').

Outputs

Outputs of a module can be defined as follows

output "my_output" {
    value = "my_output_value"
}

For root modules (i.e. the directory in which you run terraform init/plan/apply), the output values are printed to stdout. For nested modules you can refer to the output of that modules as e.g. module.my_module.my_output.

---

Iteration

over resources

Count

The count meta-argument expects an integer as its value, specifying the number of copies of an object you want to create. You can refer to the current iteration within the scope of the resource/data source/module you are creating by means of a count object, which has a index attribute.

Example:

resource "cloud_resource" "my_resource" {
    count = 5
    name = "my-resource-${count.index}"
}

For each

A more flexible iteration meta-argument is for_each which takes as an input a set, or a map (if you want to use a list, have a look at the toset function). You can refer to the keys and/or values in that set/map within the scope of the resource/data source/module you are creating with the each keyword.

Example:

resource "cloud_resource" "my_resource" {
    for_each = {key1: "value1", key2: "value2"}
    name = each.key
    value = each.value
}

over iterables

For

You can create new iterables (lists, sets, objects, maps) by iterating over another iterable with the for keyword. E.g. [for s in var.list : upper(s)] creates a new list by uppercasing the strings s in the variable list.

For other use cases, have a look at the official HCL docs.

---

Conditional expressions

Terraform allows you to conditionally define or instantiate things with its ternary operator. The expression condition? true_val: false_val evaluates to true_val if condition is an expression that evaluates to true, otherwise it evaluates to false_val.

A very common usage pattern of the ternary operator in terraform is to combine it with the count meta-argument to optionally instantiate a resource/data source/module based on some condition:

resource "cloud_resource" "my_resource" {
    count = condition? 1: 0
}

This snippet only instantiates a cloud_resource when the condition evaluates to true.