Your First AWS Lambda Function

Lambda is the simplest cloud compute. From zero to a callable function.

Step 1: Write the function

def handler(event, context):
    return {"statusCode": 200, "body": "hello"}

Save as app.py.

Step 2: Deploy via SAM

• Install SAM CLI; create a SAM template.yaml referencing app.py.
• sam build && sam deploy --guided, deploys the function + API Gateway.

Step 3: Invoke

Invocation is where the abstraction proves itself. The same function answers HTTP, scheduled events, and queue triggers depending on the wiring.

• HTTP. curl https://<your-api-url> hits the API Gateway endpoint and returns 'hello'.
• Cold start. First call after idle takes longer (200ms to 2s); subsequent calls are warm.
• Direct invoke. aws lambda invoke --function-name myfn out.json bypasses API Gateway for testing.
• Logs. Every invocation lands in CloudWatch Logs under /aws/lambda/<function-name> within seconds.

Step 4: Logs and monitoring

CloudWatch is the default observability surface for Lambda. Built-in metrics cover invocation health; custom metrics cover the business logic.

• Logs. Every invocation logged with request ID; filter-pattern queries find the failing invocations fast.
• Built-in metrics. Invocations, Duration, Errors, Throttles, IteratorAge for streams; the four-row dashboard.
• Custom metrics. Emit via boto3 CloudWatch PutMetricData or via embedded metric format in logs.
• Alarms. Error-rate alarm and p99 duration alarm at minimum; the equivalent of the four golden signals for Lambda.

Antipatterns

• Lambda without timeout config. Default 3s; many use cases need more.
• Hard-coding secrets. Use Secrets Manager.
• Default memory size. Tune for your workload.

What to do this week

Three moves. (1) Run the tutorial end-to-end on your own laptop / sandbox. (2) Apply the pattern to one production workload. (3) Document the variations you needed; share with the team.