Nora K.
Sample Least privilege done properly
Every function gets exactly the permissions it needs. I rewrote our team's SAM templates using the policy template examples.
Serverless on AWS: Lambda, API Gateway and DynamoDB
Build and deploy a real HTTP API with AWS SAM and Python, model data in DynamoDB, and handle permissions, errors, observability and cost
Created by Mateo Silva
- Updated September 2026
- English
- Intermediate
What you will learn
- Build and deploy Lambda functions behind an API Gateway HTTP API with AWS SAM
- Design a DynamoDB table from access patterns and query it efficiently
- Grant each function least privilege permissions
- Process work asynchronously with SQS, EventBridge and dead letter queues
- Add structured logs, metrics and traces to serverless code
- Estimate costs and decide when serverless is or is not a good fit
Course content
Serverless foundations
Data modelling with DynamoDB
- Video: DynamoDB in one lesson: keys, items and capacity 11 min
- Video: Starting from access patterns 12 min
- Article: Single table design for the bookings API 13 min
- Video: Conditional writes and transactions 11 min
- Video: Querying, pagination and filters 9 min
Authentication and API design
- Article: JWT authorisers on HTTP APIs 11 min
- Video: CORS, custom domains and stages 9 min
- Video: Throttling and protecting the API 9 min
- Article: Versioned and documented endpoints 8 min
Asynchronous work
- Video: Queues with SQS and Lambda 11 min
- Video: Dead letter queues and retries 10 min
- Article: Events with EventBridge 10 min
- Video: Idempotent handlers 10 min
Observability and testing
- Video: Structured logs, metrics and traces with Powertools 12 min
- Article: Alarms that point to the problem 9 min
- Video: Testing serverless code 11 min
- Video: Performance: cold starts, memory and SnapStart 10 min
Shipping and running it
- Article: CI/CD for SAM applications 11 min
- Video: Cost model: when serverless wins and when it does not 10 min
- Video: Tearing it all down 6 min
- Quiz: Final check 6 min
Requirements
- Comfortable with Python
- Basic AWS knowledge: IAM, regions, the console
- An AWS account and the AWS CLI configured
Description
Serverless lets you run code without managing servers, scale to zero when idle and pay per request. It also brings its own problems: permissions spread across many functions, data modelling that does not look like SQL, cold starts and debugging across services. This course builds one real application and tackles each of those honestly.
The project
A bookings API for a small workshop business: customers list available sessions, make bookings and receive a confirmation; staff see bookings per day. You build it with:
- AWS Lambda functions in Python
- Amazon API Gateway HTTP APIs with JWT authorisation
- Amazon DynamoDB with a single table design built from access patterns
- Amazon SQS and EventBridge for asynchronous work
- AWS SAM to define, test locally and deploy everything as code
What makes this course different
We spend real time on the parts tutorials skip: least privilege IAM for every function, idempotent handlers, retries and dead letter queues, structured logging and tracing with Powertools for AWS Lambda, and a cost model you can explain to a manager. You will also learn when serverless is the wrong choice.
All resources fit comfortably in a small monthly budget, and the final lesson removes everything with one command.
Who this course is for
- Backend developers building APIs on AWS
- Teams moving small services off always on servers
- Engineers who tried a Lambda tutorial and want production habits
Instructor
Mateo Silva
Builds serverless and Kubernetes platforms on AWS. Teaches Lambda, CDK and Helm with real projects
- 4.4 instructor rating
- 21 reviews
- 4,470 students
- 3 courses
- Sample
I am a platform engineer who moves between two worlds: serverless applications on AWS Lambda and container platforms on Kubernetes. Teams usually need both, and I teach the trade offs honestly.
Each of my courses is built around one project that grows from a single file to something you could hand to a team: an HTTP API, a CDK app with tests, a Helm chart with sane defaults. I care about the parts tutorials skip, such as permissions, configuration per environment, rollbacks and cost.
I write infrastructure in TypeScript and YAML most days and I will show you the mistakes I made so you can skip them.
Student reviews
Matt R.
Sample Powertools section is excellent
Structured logs, metrics and tracing with almost no extra code. Following one request across services was very satisfying.
Ahmed S.
Sample DynamoDB modelling finally clicked
Starting from access patterns before designing the table was the missing step for me. The bookings example is realistic.
Ifeoma C.
Sample The race condition demo was great
Watching parallel requests try to grab the last seat, then fixing it with a conditional write, made the concept stick.
Elena P.
Sample Very practical, a lot to absorb
The single table design lesson needed a second watch, but it is the best explanation I have seen.
Yara L.
Sample Honest about costs
I liked the comparison with containers at different traffic levels. Not every course admits when serverless is the wrong choice.
Kevin T.
Sample Good but Python only
Content is strong. Our team uses TypeScript, so I had to translate the examples myself.
Keep exploring
All AWS courses
More in AWS
$64.99
Enroll now