Developers find it challenging to write Amazon Web Services (AWS) CloudFormation templates in JSON or YAML in a timely manner. Defining your infrastructure using JSON or YAML can be challenging when you have a complex solution. Using the AWS Cloud Development Kit (CDK) with your preferred language, you can deploy your services with just a few lines of code. The AWS CDK allows you to use commands to build and generate YAML or JSON code that can be used to deploy CloudFormation stacks directly via the console or even to enable infrastructure as code implementation.
In this example, you can quickly build a serverless backend application with a few lines of python code via Cloud Development Kit (CDK). Depending on the use case, we can send our application's point of entry to make a request and be able to store different types of data (i.g., images or metadata). The stack will deploy: 1) an API Gateway endpoint, to allow the user to send request(s). 2) AWS Lambda to process the request(s). 3) Amazon S3 bucket to store images or files uploaded. 4) DynamoDB table to store your metadata. 5) Cognito User Pool to authenticate your users and secure your API Gateway endpoint. 6) AWS IAM roles with permissions to access different services.
- Amazon API Gateway - Entry point for your backend
- AWS Lambda - Process JSON POST request
- Amazon S3 - Bucket to store images or files
- Amazon DynamoDB - Table for to store metadata
- Amazon Cognito - Secure your backend with User Pool
- AWS IAM - Creates permissions
- An active AWS account
- AWS Command Line Interface (AWS CLI) (Install and configure)
- AWS CDK Toolkit, (Install and configure)
- Postman account or similar tool
- This example assumes the uploaded image/file has a maximum size of 2MB without compression. (Lambda has a maximum 6MB payload limit.)
- This example assumes the uploaded image/file is in a base64 format.
- No UI to authenticate user or upload image/file.
Since the stack's definition requires the user to provide a uniqe S3 bucket name, you will need to pass in your desired bucket name as a parameter. Use the command below to deploy the CDK stack:
cdk deploy ServerlessBackendStack --parameters uploadBucketName=globallyuniquebucketname
- Note:
- Stack deployment will fail if the bucket name exists or has been deployed before, you must choose a uniqe name!
- You need to change the stack's name if it's not "ServerlessBackendStack."
There are no manual steps involved in deploying this stack, however, in this example, Lambda code expects:
-
JSON POST request must have the following keys:
{ userid : "string" photo: "string - base64 format" }
-
Requires authenticated API calls. This solution assumes you have a frontend with a login UI, which allows you to configure the deployed Cognito with. If you do not have a UI and would like to test the API invoke URL via the CLI, you can follow the following instructions:
-
Create Cognito User
Once the CloudFormation template is finished deploying, you can follow the next steps to create your user:
-
Navigate to CloudFormation in the console
-
Select the deployed stack and navigate to the “Resources” tab
-
Find the created User Pool by searching AWS::Cognito::UserPool and 4. click on Physical ID for UserPool
-
Navigate to “Users” and Click “create user”
-
Complete creating the user and save the username and password for later use
-
Go to your terminal to run some of the AWS CLI commands below:
The password for the new user needs to be permanent. To do so, run the following command (Change the value of user-pool-id to the deployed Cognito User Pool ID. Change the username and password to your created values):
aws cognito-idp admin-set-user-password --user-pool-id us-east-1_xxxxxx --username usernameExample --password passwordExample --permanentNext, the IdToken of the account needs to be retrieved. To do so, run the following command (Change the value of client-id to the deployed Cognito User Pool Client ID. Change the username and password to your created values):
aws cognito-idp initiate-auth --region us-east-1Example --auth-flow USER_PASSWORD_AUTH --client-id XXXXXXXXXXXXXXXXXX --auth-parameters USERNAME=usernameExample,PASSWORD=passwordExample- Note: Ensure the region name matches where the stack was deployed in.
For example, if you have deployed your stack in us-east-1, then your region would be "us-east-1"
This command will return multiple keys. Save the IdToken for later use
- Note: Ensure the region name matches where the stack was deployed in.
For example, if you have deployed your stack in us-east-1, then your region would be "us-east-1"
-
-
Use Postman to test your API endpoint
- Note: Postman is developed by a third-party company. It isn't developed or supported by Amazon Web Services (AWS). To learn more about using Postman, or for assistance with issues related to Postman, see the Support center on the Postman website.
- Navigate to the Resources section of the CloudFormation in the console
- Find the created API Gateway by searching AWS::ApiGateway::RestApi and click on Physical ID
- Selected “Stages” on the menu on the left, then select “prod.” 4. Save the Invoke URL for later use
- Navigate to Postman and create an account
- Create a HTTPS Post Request
- Select the POST command and paste the API Gateway URL.
- Append “/form” at the end of the URL
- Go to Headers and add a key “Authorization” with the value being the saved value IdToken
- Go to Body and select raw. Change the blue “Text” setting to “JSON”
- In the test area create a request in JSON. Image details must be in base64 format { "userid": "myUserID", "photo": "image details" }
- Click “Send” on the top right. This should return a “Success!” message in the response
You should see the uploaded object in your deployed S3 bucket, and its metadata in DynamoDB table!
-
cdk lslist all stacks in the appcdk synthemits the synthesized CloudFormation templatecdk deploydeploy this stack to your default AWS account/regioncdk diffcompare deployed stack with current statecdk docsopen CDK documentation
TODO : AÑADIR https://dev.to/aws-builders/solving-problems-with-api-gateway-and-the-python-cdk-1m6a