AWS Lambda → serverless compute service configured with trigger

• runs code in response to events + manages compute resources
• ex. create new backend services where compute resources are automatically triggered based on custom requests
• Lambda Function → code that runs on trigger
  • configure to trigger from other AWS services, on events in app, on activity from other devices
• benefits
  • no servers to manage
  • continuous/automatic scaling on trigger
  • pricing based on use → # of requests + time code takes to run

• user uploads image to source bucket (object-created event)
• s3 publishes event to lambda, passes event data as function params (source bucket name + object key)
• lambda runs function → reads object + performs function (creates thumbnail + saves to target bucket)

create Lambda Function

• read image from source bucket → resize image → store in target bucket

• create function

  • author from scratch
  • use blueprint → code templates for common AWS functions
  • container image → deploy container image on trigger

• custom execution role

  • lambda-execution-role → gives Lambda permission to read + write to S3 buckets

• networking → define VPC to make secure

  • VPC
  • subnets
  • security group

• add trigger

  • trigger function whenever new object is created in source bucket

  import boto3
  import os
  import sys
  import uuid
  from PIL import Image
  import PIL.Image
  
  s3_client = boto3.client('s3')
  
  def resize_image(image_path, resized_path):
      with Image.open(image_path) as image:
          image.thumbnail((128, 128))
          image.save(resized_path)
  
  def handler(event, context):
      for record in event['Records']:
          bucket = record['s3']['bucket']['name']
          key = record['s3']['object']['key']
          download_path = '/tmp/{}{}'.format(uuid.uuid4(), key)
          upload_path = '/tmp/resized-{}'.format(key)
  
          s3_client.download_file(bucket, key, download_path)
          resize_image(download_path, upload_path)
          s3_client.upload_file(upload_path, '{}-resized'.format(bucket), key)
  

• monitoring

  • cloudwatch → log stream

knowledge check

1. most efficient way to trigger lambda function for multiple S3 buckets
   1. configure event notifications for each bucket to the same lambda function
2. key component when creating lambda function
   1. function blueprints → pre-configured templates for common use cases
3. memory setting when configuring lambda function
   1. amount of RAM allocated for function - impacts CPU allotted to run function too
4. amazon cloudwatch
   1. used to monitor lambda functions
   2. collects and tracks metrics, logs, alarms
5. maximum execution time for lambda function
   1. 15 min