DataLane

All stacks · Warehouses & analytics

Amazon Redshift

RA3, Spectrum, concurrency scaling, and the AWS warehouse.

Amazon Redshift cover

Related reading

About Amazon Redshift

Amazon Redshift is still the warehouse on many AWS estates: RA3 managed storage, Spectrum on S3, concurrency scaling, and a distribution/sort-key model that punishes you if you pretend it is Snowflake.

These notes sit next to the Redshift vs Snowflake comparison. Learn sort keys, WLM, and why SELECT * on a 400-column fact is a Spectrum tax.

What you'll learn here

  • RA3 vs DC2 and what “managed storage” actually bills
  • Distribution styles, sort keys, and vacuum/analyze habits
  • Spectrum, federated query, and when to stop copying S3 into Redshift
  • Concurrency scaling and the surprise bill

Frequently asked questions

Should a new AWS shop pick Redshift or Snowflake?

Redshift if the org is AWS-only, IAM-centric, and already pays for Spectrum/S3. Snowflake if you want multi-cloud, sharing, and less WLM archaeology. Both can query Iceberg now — the lock-in is people and IAM, not files.

Why is my Redshift scan huge?

Missing or stale sort keys, SELECT *, and Spectrum over unpartitioned Parquet. EXPLAIN and STL tables tell you. Adding nodes to hide a bad sort key is how the bill grows.

Is Redshift Serverless “no ops”?

Fewer cluster tickets, same SQL sins. Serverless still meters RPU. A bad dashboard on a 10 TB fact will find the credit card.

New Amazon Redshift posts, straight to your inbox

One email a week with our latest tutorials. No spam.

Newsletter signup is not live yet. Use the contact form if you want to be notified.

↑↓ navigate openesc close