Skip to main content

My Wild Ride through Backend Deployment: Nest.js on AWS

Hello, fellow developers! Today, I'm going to share my recent endeavor deploying a Nest.js backend on AWS. It was quite a journey, filled with both unexpected hurdles and valuable lessons. I hope that by sharing my experience, it might help you on your deployment journeys.

My Initial Try with Elastic Beanstalk​

Starting off, I decided to go with AWS Elastic Beanstalk. If you're unfamiliar with it, Elastic Beanstalk is an AWS service that automatically handles deployment details such as capacity provisioning, load balancing, and automatic scaling. It seemed like the perfect choice, but I ran into a major roadblock. My application uses Prisma,an ORM, and this led me straight into the infamous "Prisma client failed to generate" error. This error usually happens when Prisma Client is unable to execute the prisma generate command during the postinstall phase, which was the case for me.

I decided to hit the books and find a solution, and I found that Prisma has a guide to solve this issue. However, I found that the solutions provided in the guide didn't quite work for me. After what felt like an eternity wrestling with Elastic Beanstalk and Prisma, I decided to look for a different solution.

Feeling somewhat defeated but determined, I moved onto AWS CodePipeline and ECB (Elastic Container Service). These services provide continuous integration and continuous delivery (CI/CD) and container orchestration, respectively. However, the setup of these services felt like solving a complex jigsaw puzzle without the picture on the box. As a newcomer to this space, I found it difficult to connect all the pieces. Therefore, I decided to seek an alternative.

Discovering Docker​

At this point, I stumbled upon Docker. Docker is a brilliant tool that allows developers to package applications with all the parts it needs, like libraries and other dependencies, and ship it all out as one package. This feature, known as containerization, ensures that the application runs the same regardless of the environment it's running on.

What was fantastic about Docker was that it provided me with the flexibility to move from development to production without worrying about environment setup. The lightweight nature of Docker containers means I could deploy multiple instances of my application on the same machine, saving on costs without compromising on performance.

The Detour: Google Cloud Platform's Cloud Run​

With my newly Dockerized app, I felt a new sense of optimism and decided to try Google Cloud Platform's Cloud Run. Cloud Run is designed to run stateless containers, which seemed like a perfect match for my Dockerized application. However, the process to update my application on Cloud Run was far from straightforward. I had to navigate through various services, and updating the code became a long, winding journey. So, I thought it was time to return to AWS, where I'd started.

The Breakthrough: AWS EC2​

I finally decided to give AWS EC2 a shot. EC2 stands for Elastic Compute Cloud, and it provides secure and resizable computing capacity in the cloud. It's designed to make web-scale cloud computing more approachable, and for someone still finding their feet in this space, it was exactly what I needed.

Deploying my application on EC2 felt like reaching the finish line. I won't lie, the initial setup was a bit of a learning curve. There were decisions to be made about