developers don’t have to manage servers anymore they just deploy code, functions
Serverless
All serverless services
Virtual functions limited by time tthat run on-demand
Lambda
First 1,000,000 requests are free, $0.20 per 1 million requests thereafter
Lambda
Max execution time of a lambda function
15 minutes
A code that you write and attach to CloudFront distributions that runs close to your users to minimize latency
Edge Function
Lightweight functions written in JavaScript for high-scale, latency-sensitive CDN customizations that allow millions of requests/second with Sub-ms startup times
CloudFront Functions
functions written in NodeJS or Python that scales to 1000s of requests/second
Lambda@Edge
Does Lambda@Edge trigger from Origin Request/Response
Yes
CloudFront Functions vs. Lambda@Edge
CloudFront Functions
CloudFront Functions vs. Lambda@Edge
Lambda@Edge
By default, is your Lambda function
launched inside your own VPC?
No
Lambda must be configured to launch in the VPC by defining the VPC ID, the Subnets and the Security Groups
True
Lambda will create an ENI (Elastic Network Interface) in your subnets to be launched in your VPC
True
IImproves Lambda functions directly accessing your DB by by pooling and sharing DB
connections
RDS Proxy
If using Lambda with RDS proxy it must be deployed in your VPC
True
Allows you to process data events from within a RDS for PostgreSQL and Aurora MySQL DB by allowing outbound traffic, and required permissions
Lambda
Notifications that tells information about the DB
instance itself (created, stopped, start, …)
RDS Event Notifications
NoSQL fully managed Database-as-a-Service (DBaaS) product available within AWS. It is highly resilient across multiple AZ in a region or globally
DynamoDB
The capacity mode used with an unknown, unpredictable load or low admin overhead on a DynamoDB table
On-Demand
The capacity mode in DynamoDB where the RCU and WCU are set on a per-table basis
Provisioned mode
Most flexible way to get data from a DynamoDB table but the least efficient by moving through the table and consuming the capacity of every item
Scan
Reading DynamoDB data which is directed at a random node across the AZs. Possible to return with stale data if a node is checked before replication completes
Eventually Consistent reads
Reading DynamoDB data that is directly connected to the leader node which always returns the most up-to-date copy of data
Strongly Consistent reads