What we do
Tools
Industries
Department
How we workLet's talk →
Home/AWS Developer
AWS Development

Hiring an AWS Developer When the Bill Grew and Nobody Knows Why

Most businesses that search for an AWS developer already have an AWS account. Someone built it, the app works, and every month the invoice is a little higher. Before you hire for the next feature, it helps to know what you're already paying for.

One monthly AWS bill, and what each line is often paying for

Open Cost Explorer and group by service, then by usage type. Lines like these turn up again and again in accounts that grew without anyone watching cost.

Example only: a small web app and its database
EC2 instances (sized for a launch spike that never came) $410.00
RDS database (includes a staging copy running around the clock) $290.00
NAT gateway hours and data processed (S3 traffic routed through it instead of a free gateway endpoint) $180.00
CloudWatch Logs (log groups left on the default, which never expires) $95.00
EBS volumes and snapshots (left behind by servers deleted long ago) $70.00
Public IPv4 addresses (AWS charges for each one, attached or not) $22.00
Monthly total $1,067.00

Only the first two lines are the app doing its job. The rest is set-up debt: each one is a known fix once someone looks, and Compute Optimizer and cost allocation tags make the next review faster.

Clicked together in the console, or defined in code?

This decides how safely anyone can change the setup later, including us.

Console

Built by hand, one screen at a time

  • Quick to start, and common in accounts set up by one person.
  • Nobody can list everything that exists without hunting through each region and service.
  • A change made in a hurry is hard to review and harder to reverse.
Code

Written in CloudFormation, CDK or Terraform

  • Every resource is in files the business keeps in version control.
  • Changes are reviewed before they apply, and a staging copy is the same files with different settings.
  • A new developer reads the files instead of reverse-engineering the console.

You don't have to convert everything at once. We usually bring the parts we're about to change into code first, and leave the stable parts until there's a reason to touch them.

Ask these before an AWS developer touches your account

AWS has a long list of services and most developers know a handful of them deeply. These questions show which handful.

  • Which of the services we use have you run in production?

    Lambda, ECS, RDS, S3 and SQS each have their own failure modes and cost traps. General AWS experience isn't the same as having run yours.

  • How would you find what is driving our bill?

    A good answer names Cost Explorer by usage type, tags per environment and a budget alert, not a guess.

  • Will the infrastructure you build be in code we keep?

    If the answer is "I'll set it up in the console", the next developer starts from zero again.

  • How will you get access, and how will we take it back?

    IAM roles or users per person, never the root account or a shared key, so access ends cleanly when the work does.

  • How will we know when something fails?

    CloudWatch alarms on the things customers notice, and a budget alert so a cost spike is an email, not an invoice.

AWS work, DevOps work and the code that runs on it

AWS developer work is about the services themselves: how the app uses Lambda, queues, storage and the database, and what that costs. Deployment pipelines and release process are closer to a DevOps engineer, and the application code is backend developer work. Slow or oversized databases often need a database developer as much as an AWS change. In practice one project touches several of these, and n-frames covers them with its own team.

The quiet risk in a production AWS account is a change made straight in the console with no record of who made it or why. A security group opened for a quick test, or a database setting changed during an outage, stays that way until it causes the next one. Putting changes through code and review closes that gap. If your company is on Microsoft instead, see Azure developer.

Let's talk

An AWS account you can explain, line by line

n-frames reviews the account first, so the quote starts from what's really there: what each service does, what's oversized, and what's missing alarms. Then we build the new piece in code your team keeps, sized for the load you actually have.

Cut the bill, then build the next piece →

Questions people ask

Do you build new AWS infrastructure or also fix existing setups?

Both. A lot of what reaches us is an AWS setup that grew messy or over-provisioned over time, and tidying that is usually cheaper than rebuilding from scratch.

Is this different from hiring a DevOps engineer?

AWS developer work is about the AWS services themselves. DevOps is broader: the deployment pipeline and release process, whichever cloud you're on. They overlap, so describe the actual problem and we'll put the right people on it.

What if we're already committed to AWS but something's not working well?

Say which it is: cost, speed or reliability. We diagnose it against the setup you have rather than suggesting a rebuild by default.

Can you check whether we are over-provisioned before quoting new work?

Yes. A lot of AWS infrastructure we inherit is sized for a load that never arrived. Looking at that first is part of the free quote, so the new work doesn't sit on top of capacity nobody uses.

Let's talk

Tell us the one thing your team does manually that eats up time. We read every message and reply within a day.