DEV Community: Ezile Mdodana

Day 27/30: Your test server does not need to work the night shift

Ezile Mdodana — Wed, 26 Aug 2026 09:22:51 +0000

Development and testing environments are often left running all night, all weekend and sometimes long after the project ends.

If nobody uses them outside working hours, schedule them to stop where appropriate.

Also remove abandoned databases, storage volumes, snapshots and IP addresses.

AWS charges for resources, not intentions.

Small unused resources have a remarkable ability to form a large monthly bill when nobody is watching.

Day 26/30: Your database password is not source code

Ezile Mdodana — Tue, 25 Aug 2026 06:13:06 +0000

Passwords, API keys and tokens should not be hard-coded into applications or committed to Git.

AWS Secrets Manager can securely store and manage sensitive values. It can also support automated rotation for compatible services.

The application retrieves the secret when required instead of carrying it around in the code.

If your password is sitting in a repository called production-final-real, it is no longer much of a secret.

Protect credentials separately from the application.

Day 25/30: “I clicked around until it worked” is not documentation

Ezile Mdodana — Mon, 24 Aug 2026 07:12:07 +0000

Infrastructure as Code lets you define cloud resources using code.

Tools such as AWS CloudFormation and the AWS CDK can create repeatable, reviewable environments.

This means development, testing and production can follow the same blueprint.

It also reduces the mystery of the one server that only works because someone changed a setting three years ago.

If infrastructure matters, it should be reproducible—not dependent on Trevor remembering what he clicked.

Day 24/30: Containers are portable lunchboxes for applications

Ezile Mdodana — Fri, 21 Aug 2026 06:08:20 +0000

They package code and its dependencies so the application can run consistently across environments.

AWS provides services such as Amazon ECS and Amazon EKS for running containers.

ECS offers AWS-native container orchestration.

EKS provides managed Kubernetes.

Kubernetes is powerful, but power comes with complexity.

Do not hire an orchestra to play one triangle.

Choose the simplest service that meets the workload’s actual requirements.

Day 23/30: SNS, SQS and EventBridge walk into an architecture meeting

Ezile Mdodana — Thu, 20 Aug 2026 07:15:10 +0000

Amazon SNS says: “I will announce the message to everyone subscribed.”

Amazon SQS says: “I will hold the message until someone processes it.”

Amazon EventBridge says: “I will examine the event and route it according to the rules.”

All three move messages, but they solve different problems.

Choosing an AWS service because its name appeared first in a tutorial is not architecture.

Start with how information must flow.

Day 22/30: Not every task must happen immediately

Ezile Mdodana — Wed, 19 Aug 2026 06:55:11 +0000

Imagine an online shop where every order must wait for the invoice, email and stock update to finish before the customer receives confirmation.

One slow process delays everything.

Amazon SQS allows applications to place messages in a queue for other components to process independently.

This reduces tight dependencies and helps systems handle traffic spikes.

A queue is the application saying:

“I received the work. I will process it properly.”

Much healthier than everyone shouting at every service simultaneously.

Day 21/30: Two questions every business should answer

Ezile Mdodana — Tue, 18 Aug 2026 06:15:09 +0000

How long can the system be unavailable?

How much data can we afford to lose?

The first answer helps define your Recovery Time Objective: how quickly service must return.

The second helps define your Recovery Point Objective: how far back your recoverable data may be.

“Immediately” and “nothing” are popular answers—until the cost of achieving them arrives.

Recovery requirements must be based on business impact, not optimism.

Day 20/30: Multi-AZ and Multi-Region are not the same thing

Ezile Mdodana — Mon, 17 Aug 2026 07:50:55 +0000

Multi-AZ distributes a workload across Availability Zones within one AWS Region.

Multi-Region uses infrastructure in more than one Region.

Multi-Region can provide stronger resilience, but it also adds cost, data considerations and operational complexity.

Not every application needs it.

Do not build a second airport when your business currently needs a backup generator.

Start with the actual business risk, then choose the architecture that matches it.

Day 19/30: A backup you have never restored is a motivational theory

Ezile Mdodana — Fri, 14 Aug 2026 14:42:09 +0000

Creating backups is only the first step.

You must also know:

What is being backed up?
How often?
How long is it r
etained?
Who can access it?
Can it actually be restored?

Test the recovery process regularly.

Discovering that a backup is incomplete during a real incident is like testing the spare tyre after the tyre has already burst.

Backups are not about storing data. They are about recovering it.

Day 18/30: The internet also needs directions

Ezile Mdodana — Thu, 13 Aug 2026 14:07:06 +0000

People remember names such as example.com.

Computers communicate using IP addresses.

Amazon Route 53 is a DNS service that helps direct users to the correct application or resource.

It can also use routing policies and health checks to influence where traffic goes.

DNS is like a GPS for your application.

When configured correctly, nobody notices it.

When configured incorrectly, the entire company suddenly becomes interested in DNS.

Day 17/30: Distance still matters on the internet

Ezile Mdodana — Wed, 12 Aug 2026 06:55:06 +0000

If your users are far from where content is hosted, every request has further to travel.

Amazon CloudFront caches content closer to users through a global network of edge locations.

This can improve performance and reduce repeated requests to your origin server.

Think of it as placing stock closer to customers instead of sending every order back to the main warehouse.

Your users do not care how impressive the architecture diagram is. They care that the page loads quickly.

Day 16/30: A VPC is your private neighbourhood in AWS

Ezile Mdodana — Tue, 11 Aug 2026 12:18:15 +0000

An Amazon Virtual Private Cloud allows you to create an isolated network for your AWS resources.

Inside it, you can define subnets, route
s, gateways and security controls.

Public-facing components can live in public subnets. Databases and internal services usually belong somewhere less exposed.

Putting every resource directly on the internet is not “keeping the architecture simple.”

It is leaving every door open because keys are complicated.

Design the network around how traffic should flow.