When we first decided to move away from AWS, we knew it wouldn’t be easy. What we didn’t expect was the biggest challenge: confronting just how much we’d come to rely on AWS’s “off-the-shelf” services.
Over the years, we’d built our architecture around whatever AWS made simple for us. Need a DB? Use RDS. Need a Data Warehouse? Hello, Redshift. DNS? Route53 is right there. Each of these services was fine, convenient, and available at the click of a button.
But here’s the trap:
We weren’t making the best technical decisions.
We were making the easiest decisions.
And those decisions came with invisible strings — both technical and financial.
The Technical Comfort Blanket
AWS makes it absurdly easy to build fast. But when you build fast with “just use what they offer” as your mantra, you lose the discipline of architectural design. You stop asking:
Is this the best tool for the job?
Will this scale in a cost-efficient way for our business model?
What happens if we want to move away from it?
We discovered that we had tied ourselves into a world where we weren’t just paying for convenience — we were paying for complacency.
The Cost Creep Reality
The other hard pill to swallow: we’d become okay with the costs.
At first, it’s small — pennies here and there for managed services. But then it balloons into monthly bills where a shocking percentage is from things we could have managed differently, more efficiently, or avoided altogether.
The ease of AWS’s services blinded us to the long-term cost implications. We stopped optimising. We stopped questioning.
The Turning Point
- Leaving AWS forced us to relearn the art of intentional architecture.
- We had to question every piece of our infrastructure.
- We learned to rely on open-source tools that we could host, optimise, and control.
- We became more aware of cost per feature and what “easy” was really costing us.
And — crucially — we stopped outsourcing architectural decisions to a hyperscaler.
What We Learned (The Hard Way)
- Convenience comes at a cost — and not just financial.
- Vendor lock-in isn’t just a buzzword; it’s a business risk.
- “Best practice” doesn’t always mean “best for your business.”
- Open-source and vendor-agnostic solutions may take more initial effort, but they give you freedom.
- AWS (or any hyperscaler) will never push you to spend less — that has to be your job.
Where We Are Now
We’ve rebuilt on a stack that gives us flexibility, visibility, and control. We’re running what we need, where we need it — on infrastructure we can move, scale, and optimise without fearing a monster bill.
Would we ever go back? Probably not.
Do we still appreciate what AWS taught us? Absolutely.
But the biggest lesson of all: stop choosing easy. Start choosing smart.
