AWS CDK - test and production environments

Last week, I used ChatGPT to write Python code that converts a CSV file uploaded to an S3 bucket into a table in an RDS (transactional database). Once the code was ready, I wanted to deploy it in a test environment before moving it to the production environment.

It turned out there is a convenient way to do that. First, we need to create environment.

env = Environment(account=os.environ["CDK_DEFAULT_ACCOUNT"], region=os.environ["CDK_DEFAULT_REGION"])

While writing this post, I noticed that in the above code line generated by ChatGPT, my AWS account ID was hardcoded, which is not a good practice and exposes sensitive data. This is a good reminder to review what AI creates. To avoid this mistake, we should read the AWS account ID and region from our local AWS account configuration files located in ~/.aws/credentials and ~/.aws/config.

When we deploy our app, AWS CDK will use our AWS account from the default or profiled AWS CLI configuration to set the CDK_DEFAULT_ACCOUNT and CDK_DEFAULT_REGION environment variables.

In the next step, we need to create two stacks for each environment.

app = cdk.App()
BudgetCsvTransformStack(app, "BudgetCsvTransformStack-Test", stage="test", env=env)
BudgetCsvTransformStack(app, "BudgetCsvTransformStack-Prod", stage="prod", env=env)

It is important to note that we are using a suffix in the stack name:

• -Test for the test environment

• -Prod for the production environment

Additionally, we define the stage argument in the BudgetCsvTransformStack constructor:

• stage="test" for the test environment

• stage="prod" for the production environment

It's important to note that the stage argument will be used in each AWS resource defined in the BudgetCsvTransformStack constructor to specify environment-specific resource names, for example.

bucket = s3.Bucket(
            self, f"CsvUploadBucket-{stage}",
            bucket_name=f"budget-csv-uploads-{stage}",
            removal_policy=RemovalPolicy.DESTROY,
            auto_delete_objects=True
        )

As the last step, we pass the env variable to the BudgetCsvTransformStack constructor using the environment defined earlier

Now, when we want to deploy our app to a specific environment, we can use a command like the following:

cdk deploy BudgetCsvTransformStack-Test

We could also pass AWS CLI profile in cdk deploy command

cdk deploy BudgetCsvTransformStack-Test --profile my-profile

Here we are telling AWS CDK to take our AWS account information from AWS CLI my-profile profile. So both CDK_DEFAULT_ACCOUNT and CDK_DEFAULT_REGION environment variables will be defined based this profile instead of default one.

Before using AWS CLI profile in cdk deploy command it is required to bootstrap cdk configuration for this specific profile by running following command

cdk bootstrap --profile my-profile

When we want to destroy the app in a specific environment, we can use a command like the following:

cdk destroy BudgetCsvTransformStack-Test

I think this is a very convenient way to separate testing and production environments.

Here you could find full code used in this blog post

https://github.com/jrsokolow/budget_planner/blob/main/budget-csv-transform/app.py

https://github.com/jrsokolow/budget_planner/blob/main/budget-csv-transform/src/budget_csv_transform_stack.py