Working with CloudFormation Templates

The CloudFormation template has several aspects, and each supported AWS resource has its own syntax and parameters that we need to follow. To keep things simple, we’ll only explore the basic anatomy of a CloudFormation template, which is enough to start working with templates for the commonly used general CloudFormation stacks. The following illustration provides a breakdown of the different sections in a CloudFormation template:

Here’s the basic skeleton code format for a CloudFormation template in JSON and YAML that covers the first basic layer of the template:

Here’s a brief explanation of the most common sections that we can see in both the JSON and YAML CloudFormation templates:

AWSTemplateFormatVersion

In this section, we specify the version of the template format we want to use, which is a date in the YYYY-MM-DD format. We can only provide a string literal as the template format and version.

Description

In this section, we can add information and comments to describe the CloudFormation template and the stack it will provision. We can only provide a string literal as the description, which must be 1024 bytes in length at most.

Metadata

In this section, we specify a single or list of YAML/JSON objects that provide the metadata-based information about the template. We can also use it to define configuration settings here, which are then retrieved using CloudFormation features, like CloudFormation scripts.

Parameters

In this section, we specify a single or list of YAML/JSON objects representing parameters that allow us to input custom values into the CloudFormation template when creating or updating a stack.

It can be beneficial to have parameters within our template, especially when updating a stack, as we only need to input a new parameter value once that can be referenced in the Resources and Outputs sections and avoid changing the value at multiple instances in the template.

Here’s an example template code snippet for a parameter defining the type of an EC2 instance:

Resources

In this section, we specify a single or list of YAML/JSON objects representing resources that we want to provision in the CloudFormation stack. This section supports the following fields:

• Logical ID: This field is a unique alphanumeric ID that references the corresponding AWS resource in a CloudFormation stack and within other sections in the template itself. The logical ID is different from the physical ID of an AWS resource, which is the actual ID of the resource.

• Type: This field represents the type of AWS resource that we want to create. The resource type field has the following format: service-provider::service-name::resource-type. For example, if we want to create a Lambda function, we’ll write AWS::Lambda::Function as the resource type.

• Properties: This field is where we can specify any additional properties for the AWS resource.

Here’s an example template code snippet for provisioning the EC2 instance and security group resources. Note that we also utilized the EC2InstanceType parameter from the parameter example.

Outputs

In this section, we specify a single or list of YAML/JSON objects representing outputs related to the stack that we can return via API-based stack calls or the AWS CloudFormation management console.

Here’s an example template code snippet for the output defining the public DNS link of the EC2 instance:

Now that we’ve explored the basic anatomy of a CloudFormation template, we can proceed to understand some of the basic features that CloudFormation offers to work with templates.

Intrinsic functions

Intrinsic functions are built-in functions provided by CloudFormation, and we can use these in our template to assign resource attributes with unavailable values until runtime. Here’s a brief description of some of the commonly used intrinsic functions in CloudFormation templates:

• !Base64 | Fn::Base64: This intrinsic function returns the Base64 representation of the input string.

• !FindInMap | Fn::FindInMap: This intrinsic function returns the values corresponding to the keys defined in the Mappings section.

• !GetAtt | Fn::GetAtt: This intrinsic function returns the value of an attribute from a resource in the template.

• !Ref|Ref: This intrinsic function returns the value of the specified parameter or resource.

• !Sub | Fn::Sub: This intrinsic function is used to substitute variables in an input string with values that you specify.

Pseudo parameters

Pseudo parameters are pre-defined parameters provided by CloudFormation that we don’t essentially need to declare in the template. We can retrieve the pseudo parameter using the Ref Intrinsic function. Here’s a brief description of some of the commonly used pseudo parameters in CloudFormation templates:

• AWS::AccountId: This pseudo parameter returns the ID of the AWS account in which the stack is currently being built.

• AWS::Region: This pseudo parameter returns the region value in which the stack is currently being built.

• AWS::StackId: This pseudo parameter returns the ID of the stack that is currently being built.

• AWS::StackName: This pseudo parameter returns the name of the stack that is currently being built.

This lesson taught us about the AWS CloudFormation templates and understand the core structure and functions of a CloudFormation template.