Question

A media company serves web assets through Amazon CloudFront with a single Amazon S3 bucket as the origin. During a recent regional S3 disruption, users received 5xx errors because CloudFront could not reach the origin. The company has already replicated all assets to a second S3 bucket in another Region. A CloudOps engineer must configure CloudFront so that requests automatically fail over to the secondary bucket when the primary origin returns specific error responses, without changing the viewer-facing domain name. What should the engineer do?

Accepted Answer

Create a CloudFront origin group containing both S3 buckets as origins, define the failover criteria to include the desired HTTP status codes, and associate the origin group with the cache behavior.

Answer

Create a Route 53 failover routing policy with health checks that point to each S3 bucket, and set the CloudFront distribution origin to the Route 53 record.

Answer

Enable S3 Cross-Region Replication and configure a Lambda@Edge function to rewrite the origin domain when it detects 5xx responses from the primary bucket.

Answer

Add both S3 buckets as separate origins with separate cache behaviors and use CloudFront weighted origin distribution to shift traffic during failures.