How I Saved $80k/Year by Moving from Airbyte Cloud to Airbyte OSS on EC2

As engineers, we’re often faced with the challenge of balancing performance with cost. It’s easy to fall into the trap of relying on out-of-the-box solutions that seem convenient at first but become financial burdens over time.

This is the story of how I slashed our daily costs from $250 to $0 by taking control of our data pipeline with an in-house Airbyte setup on AWS EC2 over the weekend.

The Problem: A Costly MySQL Connector

Here is the email I sent to Airbyte Support when the cost of Airbyte increased to 100X, causing us to run out of money that was supposed to last for a month.

It turned out that the MySQL Connector we were using was in beta at the time. When it became certified, they started charging for its usage, ranging from $10/day to $250/day.

$250/day x 365 days = $91,250/year.

🤯 It’s more than our AWS costs combined at that time. And this is just data transfer!

The Solution: Launch Our Own Airbyte!

Rather than continue paying the crazy expensive fees, I decided to take matters into my own hands. The solution? Deploying our own version of Airbyte OSS on an EC2 instance. Here’s how I did it.

Step 1: Choosing the Right EC2 Instance

First, we needed to select the right EC2 instance. Our pipeline wasn’t too heavy, so we went with a t3.medium instance. It’s cost-effective and provides a balance of compute and memory:

aws ec2 run-instances --instance-type t3.medium --image-id ami-XXXXXXXX --key-name MyKeyPair --security-group-ids sg-XXXXXXXX --subnet-id subnet-XXXXXX

Step 2: Setting Up Docker

Airbyte OSS runs on Docker, so the next step was to install Docker on our EC2 instance:

sudo apt-get update
sudo apt-get install docker.io
sudo systemctl start docker
sudo systemctl enable docker

Step 3: Deploying Airbyte OSS

With Docker installed, we were ready to deploy Airbyte:

mkdir airbyte && cd airbyte
wget <https://raw.githubusercontent.com/airbytehq/airbyte/master/{docker-compose.yaml,.env}>
docker-compose up -d

This setup pulled the latest Airbyte image and spun it up in the background. We then accessed the Airbyte dashboard via our EC2 instance’s public IP.

Here is the doc that I mainly followed to successfully install and get it up and running. Amazing job by the Airbyte community to keep the installation guidelines up to date.

The Results: Significant Cost Savings

By deploying our own Airbyte instance, I brought our daily costs back down to $0. The EC2 instance we used cost us around $20 per month, which is a mere fraction of what we were paying for the Airbyte Cloud version.

Look at these pipelines running on our server every day, transferring about 15–20 GB of data from different sources to Snowflake 🎉 for the last 10 months.

In total, we saved approximately $80,000 per year — funds that we could now allocate to other critical areas of our business.

Conclusion: The 10X Mindset

This experience taught me that being a 10X engineer isn’t just about writing efficient code — it’s about finding efficient solutions. Sometimes, that means questioning the tools you’re using and finding alternatives that better suit your needs. In this case, the alternative saved us $80,000 a year.

So, the next time you’re faced with a rising cloud bill, remember: there might be a better, cheaper way. And if there isn’t — well, you might just have to build it yourself.