Lambda

Question 1

What are the types and the differences between two types of lambda invocations?

Answer 1

Sync invocation:
- result is returned right away
- every time it is user triggered
- error handling client side

Async invocation:
- events are placed in an event queue
- lambda attempts to retry on fail (3 times)
- can define a DLQ for failed processing

Question 2

What is lambda event source mapping?

Answer 2

It is a way to integrate with other AWS services that will invoke AWS lambda synchronously every time a batch is returned from the service. Lambda keeps polling for data every time.
Compatible with Dynamo Streams, SQS, SNS and kinesis. Provides auto scaling

Question 3

What are lambda destinations?

Answer 3

It is a way to send, asynchronously succesful and failed events triggered from AWS lambda. Can act as a DLQ.

Question 4

What is a lambda edge function and when you should use it?

Answer 4

Code that you write and attach to cloudfront. Should be used when you need to minimize latency.

Question 5

What happens if a request is made and the limit of concurrent executions is reached?

Answer 5

Each invocation over the limit will trigger a throttle, that will return 429 if the request is sync and retry automatically and go to dlq if async.

Question 6

What is provisioned concurrency and when to use it?

Answer 6

It is a way to prevent cold start from happening, by leveraging lambda functions before they are invoked, using ASG

Question 7

What are lambda versions and aliases?

Answer 7

A lambda version is created when the function is published and it is immutable. An alias is a pointer to a lambda version, so we can call $LATEST with a DEV alias, v1 with a PROD alias, etc, etc. This way it is easier to provide URLs to the lambdas