Real-world scenario
A payment processor streams transactions to Lambda. Functions under 1000 ms detect anomalies without running a 24/7 server.
By the end you can:
Follow the request/data flow through the architecture.
Work through each concept and mark it understood. This unlocks the knowledge check.
Run code in response to events without provisioning servers; billed per request + duration.
Why it matters
Scales to zero and to thousands automatically — no idle cost, no patching.
When to use it
Use for event-driven glue, APIs, and spiky workloads. Avoid for long (>15 min) or very high-throughput steady jobs where Fargate is cheaper.
Example
def handler(event, context):
# Triggered by API Gateway / S3 / EventBridge
name = event.get("queryStringParameters", {}).get("name", "human")
return {"statusCode": 200, "body": f"*bleep* hello, {name}"}Gotcha (exam trap)
Cold starts add latency; memory and CPU are coupled (more memory = more CPU). The 15-minute hard timeout is a frequent exam trap.
Use the simulated AWS console below to complete each mission step. No real AWS account is required.
Create a Lambda function named "fraud-detector", runtime python3.12, handler index.handler, role arn:aws:iam::123456789012:role/etl-service-role, with any Python code.
Invoke "fraud-detector" with a payload like {"Records":[{"transaction_id":"txn-2001","amount_usd":25000}]}.
Bleepx skills build on each other. Continue your learning path across Query, Lab, and Cloud.