Automating Deployments with AWS Cloudformation

Task 1: Deploy a CloudFormation Stack

I began by deploying a CloudFormation stack that creates a VPC.

First, I downloaded the CloudFormation template (task1.yaml) by right-clicking the link and opened it in a Text Editor (not a Word Processor).

Looking through the file, I noticed several sections:

• The Parameters section used to prompt for inputs - here it was asking for two IP address (CIDR) ranges for defining the VPC
• The Resources section used to define the infrastructure to be deployed - defining the VPC and a Security Group
• The Outputs section used to provide selective information about resources in the stack - providing the Default Security Group for the VPC

I noted that the template was written in YAML format, commonly used for configuration files, and that the format was important, including indents and hyphens. I also learned that CloudFormation templates can be written in JSON as well.

To launch the CloudFormation stack:

1. I opened the AWS Management Console and clicked CloudFormation from the Services menu
2. Clicked Create stack then selected Upload a template file
3. Clicked Browse/Choose file and uploaded the task1.yaml template I downloaded
4. Clicked Next
5. On the Specify Details page, I named the stack "Lab"
6. I observed the Parameters section prompted for IP address ('CIDR') ranges for the VPC and Subnet with default values specified, so I left them unchanged
7. Clicked Next
8. On the Options page, I browsed through but left settings at their default values
9. Clicked Next
10. On the Review page, I saw a summary of all settings and acknowledged that custom names were being used
11. Clicked Create stack

The stack entered the CREATE_IN_PROGRESS status. I clicked the Events tab and scrolled through the listing, seeing the activities performed by CloudFormation in reverse order, such as starting and completing resource creation.

I also checked the Resources tab, which showed the resources being created. I noticed that CloudFormation was determining the optimal order for resources to be created, such as creating the VPC before the subnet.

I waited until the status changed to CREATE_COMPLETE, refreshing occasionally to update the display.

Task 2: Add an Amazon S3 Bucket to the Stack

For this task, I needed to edit the CloudFormation template to add an Amazon S3 bucket and then update the stack.

I opened the task1.yaml file I downloaded earlier and referred to the Amazon S3 Template Snippets documentation page for assistance.

Looking at the YAML example, I added the following code under the Resources header in the template:

I was careful with the indentation, using two spaces for each indent, since YAML is very particular about formatting. I didn't add any Properties since they weren't required for this basic bucket definition.

To update the stack:

1. In the CloudFormation console, I selected the Lab stack
2. Clicked Update
3. Chose Replace current template
4. Selected Upload a template file and browsed to my modified task1.yaml file
5. Clicked Next through the Specify stack details page
6. Clicked Next on the Configure stack options page
7. Reviewed the preview of changes at the bottom of the page, which showed CloudFormation would Add an Amazon S3 bucket while all other resources would remain unchanged
8. Clicked Update stack

After about a minute, the stack status changed from UPDATE_IN_PROGRESS to UPDATE_COMPLETE. I clicked the Resources tab and saw the bucket displayed in the list. CloudFormation had assigned it a random name to avoid conflicts with existing buckets.

I also verified the bucket creation by checking the S3 console.

Task 3: Add an Amazon EC2 Instance to the Stack

This task was more complex as it required adding an Amazon EC2 instance to the template. First, I needed to add a special parameter for the Amazon Machine Image (AMI).

I updated the template by adding these lines in the Parameters section:

This parameter uses the AWS Systems Manager Parameter Store to retrieve the latest AMI for the stack's region, making it easy to deploy stacks in different regions without manually specifying an AMI ID for every region.

I learned that when writing CloudFormation templates, I could refer to other resources using the !Ref keyword. For example, I saw in the template how a VPC was defined and then referenced within the Route Table definition:

Using the AWS::EC2::Instance documentation page, I added the EC2 instance definition to my template:

I made sure to include only the five Properties listed in the requirements:

• ImageId: Referenced the AmazonLinuxAMIID parameter
• InstanceType: Set to t3.micro
• SecurityGroupIds: Referenced AppSecurityGroup (noting that it needed to be in list format)
• SubnetId: Referenced PublicSubnet
• Tags: Added the Name tag with value "App Server"

To update the stack:

1. I selected the Lab stack in the CloudFormation console
2. Clicked Update
3. Selected Replace current template and uploaded my revised template file
4. Clicked through the wizard pages
5. Reviewed the preview of changes showing the EC2 instance addition
6. Clicked Update stack

Once complete, I verified the instance was displayed in the Resources tab and checked the EC2 console to see my App Server.

Task 4: Delete the Stack

For the final task, I deleted the CloudFormation stack to see how it would automatically remove all resources it created.

In the CloudFormation console:

1. I selected the Lab stack
2. Clicked Delete, then clicked Delete stack at the prompt
3. The stack showed DELETE_IN_PROGRESS status
4. After a few minutes, the stack disappeared

I verified that the Amazon S3 bucket, Amazon EC2 instance, and the VPC had all been properly deleted.