Unboxing S3 Object Lambda
Amazon S3 Object Lambda offers a way to execute a Lambda function when someone wants to download a file (GetObject) from an S3 bucket. You can implement whatever logic you wish and return any data as the response via the WriteGetObjectResponse API. Keep in mind that the Lambda function must finish within 60 seconds and is called synchronously.
The following figure shows the needed components:
Looking for a new challenge?
Let’s look at the parts more closely:
- S3 Object Lamda Access Point: References your Lambda function and your S3 Access Point. Block Public Access is always enabled.
- Lambda function: Invoked when a GetObject request is made to the S3 Object Lambda Access Point.
- S3 Access Point: Takes care of the non GetObject requests (such as ListObjects) and serves the pre-signed URL of the original file passed into the Lambda function.
- S3 Bucket: Stores the original file (in fact, the original file could be non-existent).
In the following video, I go into more details and share code with you:
- Introducing S3 Object Lambda
- Use Cases
- Demo: Generating content
- Understanding Pricing
- Pitfall: Keep the Lambda function running
You pay for the following:
- S3 Get Object API (as usual)
- Lambda invocation
- Lambda GB-second
- $0.005 per GB returned via WriteGetObjectResponse
To invoke the S3 Object Lambda Access Point, run with your ARN:
aws s3api get-object --bucket arn:aws:s3-object-lambda:us-east-1:123456789123:accesspoint/s3-object-lambda --key file.txt outfile.txt