Dynamodb Lambda Case

Thank you for reading :))

Local Development

For local development you have to create an .env file which is given in example.env file you should enter your own attributes. Also for testing you should use 8080 port. For testing, you can use "test.sh" script that is a basic script that puts 100 item in your db.

Testing

For testing if your port is 8080 you can use these commands:

Also, I recommend to use jq for checking response.

Get Item

curl localhost:8080/picus/get/{key} | jq

Post Item

curl -X POST -d '{"id": "Meaning_Of_Life", "data": "42"}' localhost:8080/picus/put

Get All List

curl localhost:8080/picus/list | jq

Deleting Item

curl -X DELETE localhost:8080/picus/{key}

For Accessing Deployment

You should replace IP with the localhost:8080

Table of Contents

1. Introduction
2. Create DynamoDB
3. Http Endpoint
4. Lambda Function and IAM Settings
5. Dockerfile
6. Build and Testing
7. Deployment
8. Mistakes and Improvements

Introduction

There is small http application's deployment and architecture. Also I added the lamba function's IAM setting.

I try to draw basic Architecture.

I give full access over the DynamoDB. So we can read/write to DynamoDB.

There are two action file. For saving resource the deployment one is only triggered after the merge.

For every request the test and build one runs.

Create DynamoDB

I used the AWS UI to create DynamoDB. For the key, I chose id which is of string type. You can choose any db name pass the env. variable

Http Endpoint

For function except DELETE from Lambda you can find under the in the Util directory.

My application makes a request to the Lambda function, which then returns the result.

Lambda Function and IAM Settings

The name of my Lambda function is delFromDyno. I have placed the code for this function in the Lambda-Code directory. I chose to use Python because it is natively supported. Also, using a different language would be a good test for me.

Dockerfile

The Dockerfile is two-staged. In the first stage, I created the baseline and built the application. In the second stage, I copied the binary from the builder stage (the first stage) and added the environment variables to the application container.

I exposed port 80, and with the entry point, I run the myapp application.

Build and Testing

It is kind of straight forward. I prepared some unit test that you can find under the test directory. I run the test and build. for each pull-request.

Deployment

For deployment, I use AWS Fargate with two instances. Deployment runs after merging the pull request.

First, I configure AWS credentials. Then, I build the Docker image and push the image to ECR. After that, I update the ECS services. I have two ECS instances with auto-scaling configured via the UI, ensuring scalability without any downtime during deployments. When I merge, AWS automatically scales up to two instances. Once the new instance is ready, AWS shuts down the old instances, ensuring seamless updates.

In case of failures, I perform rollback procedures. Additionally, I set up alerts using CloudWatch for monitoring.

Mistakes and Improvements

I found some mistakes, here is what I have identified:

• Adding a Load Balancer
  • I believe adding a Load Balancer would be better. Additionally, you can assign a static IP.
• Using root account for the task.
  • I created a new account, but I believe using the root account is not best practice. I should use a child account
• Naming the AWS resources
  • Naming resources test and test11 is not ideal for the task. I believe I should use more appropriate names