AWS Migration Journey – Part I
Greetings, dear reader! In this week’s white paper, we delve into the intricacies of our migration to AWS. Before we embark on this journey, let’s unravel the layers of our legacy public-facing ecosystems. Due to privacy concerns and the ever-looming presence of bad actors, we’ll tread lightly on the specifics of our stacks and systems. […]

Greetings, dear reader! In this week’s white paper, we delve into the intricacies of our migration to AWS.

Before we embark on this journey, let’s unravel the layers of our legacy public-facing ecosystems. Due to privacy concerns and the ever-looming presence of bad actors, we’ll tread lightly on the specifics of our stacks and systems. However, we can provide a glimpse into the vast landscape we manage.

Insanen Network and Solutions oversee the infrastructure of over 1000 clients and counting. Each client brings a unique stack, introducing various flavors into our dynamic environment. Since 2007, our commitment to accommodating this diversity has led us to adopt diverse methodologies for automating, maintaining, and scaling servers.

While the feedback from our clients has been consistently positive, claiming that everything is a walk in the park would be far from the truth.

Allow me to introduce you to Jose, our DevOps expert responsible for overseeing the migration for our client “ACME123.” Let’s hear from Jose as he breaks down a simple sample of the migration process.

“Thank you, Nick. ACME123 serves a mid-size community with 5-10k daily visitors. The technology stack for the front end involves multiple zones or a lot of micro-computers.

Before we delve further, let’s spotlight some of the pain points we faced with bare-bones VPS. Despite our best efforts to optimize servers, achieving the same speed as a t2.micro on AWS proved elusive. Whether it was the overhead from other Linux flavors in comparison to AWS Linux, this thing truly soars.

Beyond mere configuration issues, the relentless task of monitoring the same boxes for zombies or corruptions, despite our best efforts, became overwhelming.”

Wait, Jose! This is just the beginning of the conversation—let’s not spill all the beans at once, my friend.

Now, breaking it down for our readers: ACME123 operates in the clothing and music industry, specializing in the sale of various physical merchandise through their application.

Without diving too deep into the intricacies, the DevOps team’s mission was to launch a new container with a LAMP Stack Template, multiplied by three, and enable auto-scaling based on demand.

For instance, envision three containers in action. To put it in perspective, let’s take a comparison price from my personal favorite, Digital Ocean, at $8 per server with Ubuntu. This is somewhat akin to what we were privately executing.

But, team, it’s not as straightforward as merely running the template. There’s a bit of extra house cleaning involved with the newly spun servers. When a server, as per the example, started experiencing heavy traffic, the team had to replicate and scale to meet the demands. Ah, the good old days – when DevOps found themselves oscillating between craziness and occasional boredom.

Stay tuned as we unravel more layers of our AWS migration journey, addressing challenges and celebrating victories along the way.