Writing and testing Lambda code

Serverless Applications with AWS Lambda

Claudio Canales

Senior DevOps Engineer

Why testing and logging matter

Real events are messy and come in many shapes.
Testing catches errors before users do.
Logs are your "black box" for debugging.

Testing is a fire drill

One flow to remember

Core flow toolkit

Event in -> handler runs -> response out.
Use logs + test events to debug.

Lambda invocation flow

Where logs go: CloudWatch Logs

Lambda writes print() output and errors there by default.
CloudWatch Logs is AWS's log storage for services.
Log groups look like /aws/lambda/<function-name>.

Handler logs to CloudWatch Logs

The handler contract

event is the input payload.
context provides runtime metadata.
Return value becomes the function response.

Handler contract: event + context -> response

Walkthrough: event and context

def lambda_handler(event, context):
    request_id = context.aws_request_id
    time_left = context.get_remaining_time_in_millis()
    return {"statusCode": 200}

Event and context callouts

Parsing input safely

Prefer event.get() for optional fields.
Validate required inputs early.
Return clear errors for missing fields.

Parse safely and validate early

Walkthrough: validate required input

def lambda_handler(event, context):
    if not event.get("name"):
        return {"statusCode": 400, "body": "Missing: name"}
    return {"statusCode": 200, "body": f"Hello, {event.get('name')}!"}

Validation branch: 400 vs 200

Edge cases to test

Missing fields (empty JSON).
Wrong types (name is a number).
Extra fields or large payloads.

Edge case payload examples

Returning a predictable response

Keep a consistent response shape.
Use statusCode for success or errors.
JSON-encode your response body.

Response envelope: statusCode + JSON body

Walkthrough: returning JSON

import json

def lambda_handler(event, context):
    body = {"message": "ok"}
    return {"statusCode": 200, "body": json.dumps(body)}

json.dumps: dict to JSON text

Logging and tracing

Log key fields (use JSON when needed).
Include context.aws_request_id for tracing.
Avoid logging secrets or personal data.
By default, Lambda logs end up in CloudWatch Logs.

Logging best practices

Walkthrough: structured logging

import json

def lambda_handler(event, context):
    print(json.dumps({"request_id": context.aws_request_id, "event": event}))
    return {"statusCode": 200}

Structured logging JSON line

Testing with console events

Create named test events.
Test the happy path and edge cases.
Review response + logs after each run.

Console test events

A simple debug loop

Make a small code change.
Deploy, then run a test event.
Read response + CloudWatch Logs.

Debug loop cycle

Let's practice!

Serverless Applications with AWS Lambda