Amazon Redshift

Amazon Redshift

Usage Based

Accelerate your analytics with the most widely used cloud data warehouse.

Databases
OLAP Databases

Published 29 May 2026 · Last updated 27 September 2026

Scores

Popularity3/5

Widely used in AWS-heavy data teams; established but facing competition from Snowflake.

Learning Curve3/5

Similar to BigQuery; VACUUM, distribution keys, and sort keys require dedicated study.

Flexibility3/5

SQL with Redshift-specific extensions; architectural flexibility is limited to the AWS ecosystem.

Performance4/5

Columnar storage with distribution and sort keys delivers fast analytical queries at scale.

Portability2/5

AWS-specific ecosystem with Redshift-specific SQL; migration to another warehouse is costly.

About Amazon Redshift

Amazon Redshift is AWS's managed data warehouse for running analytical SQL across very large datasets, using a massively parallel, columnar architecture. Launched in 2013, it was one of the first cloud warehouses to make petabyte-scale analytics affordable, and it remains the default warehouse for many AWS-centric data teams.

It comes in two forms. Redshift Serverless removes cluster management: capacity is measured in Redshift Processing Units, scales automatically with the workload, and is billed per second while queries run, which suits intermittent or unpredictable use. Provisioned clusters let you choose node types and counts; the current RG nodes, recommended by AWS, and RA3 nodes both separate compute from managed storage, so each scales and is billed independently, and RG nodes include a built-in engine for querying data lake files.

Redshift reaches beyond its own tables. Spectrum and data lake queries read Parquet, ORC, and JSON in Amazon S3 without loading them, federated queries join live data from RDS and Aurora, zero-ETL integrations replicate operational databases in near real time, and Redshift ML trains and runs SageMaker models from SQL. Concurrency scaling adds capacity during spikes, with a daily free allowance.

The SQL dialect is PostgreSQL-based, so most Postgres syntax works, and Redshift integrates tightly with S3, Glue, Lake Formation, QuickSight, and IAM. Pricing is per RPU-hour on Serverless or per node-hour on provisioned clusters, with reserved nodes discounting steady workloads, plus managed storage per GB. Provisioned clusters still need tuning of sort and distribution keys for the best performance.

Key Features

  • Columnar MPP architecture for petabyte-scale analytics
  • Redshift Serverless with automatic scaling and per-second billing
  • RG and RA3 nodes with separate compute and managed storage
  • Spectrum and data lake queries over files in S3
  • Zero-ETL integrations and federated queries with RDS and Aurora
  • Concurrency scaling for peak workloads
  • Redshift ML with SageMaker models from SQL
  • PostgreSQL-based SQL dialect

Pros

  • Natural fit for data stacks already built on AWS
  • Serverless option removes cluster management for irregular workloads
  • PostgreSQL-style SQL eases migration from Postgres
  • Queries the S3 data lake without ETL
  • Zero-ETL replication from operational databases

Cons

  • Provisioned clusters need tuning (sort keys, distribution keys, maintenance)
  • Serverless cold starts can slow time-sensitive dashboards
  • Pricing across Serverless, provisioned, reserved, and Spectrum is hard to compare
  • Tied closely to the AWS ecosystem

Amazon Redshift Pricing

Usage Based
ServerlessContact sales
  • · $0.375 per RPU-hour (us-east-1), billed per second
  • · 60-second minimum per query
  • · Managed storage $0.024 per GB-month
  • · $300 free trial credit for 90 days
Provisioned (On-Demand)Contact sales
  • · RG (recommended) and RA3 nodes, billed per node-hour
  • · Managed storage $0.024 per GB-month
  • · Spectrum $5 per TB scanned on RA3
  • · One free hour of concurrency scaling per day
Reserved NodesContact sales
  • · 1- or 3-year node reservations
  • · Large discounts versus on-demand
  • · All, partial, or no upfront payment

Tech Stacks with Amazon Redshift

AWS Data Dashboard

Project

A minimal AWS analytics stack: store and query data in Amazon Redshift, then build dashboards and embed analytics with Amazon QuickSight, with no third-party tools required.

Data Ingestion add-on:
Data Transformation add-on:
Orchestrator add-on:

Tools Related to Amazon Redshift

Works well with Amazon Redshift(3)

Looker connects to Redshift over its Postgres-compatible JDBC interface, and LookML models generate the SQL that runs in Redshift, so metrics are defined once and computed in the warehouse.

Redshift works with Superset through the sqlalchemy-redshift driver, giving a free, open-source dashboard layer over warehouse tables without extracting data.

Tableau connects to Redshift through its native connector with IAM or SSO authentication, querying live or building extracts for dashboards.

Integrates with Amazon Redshift(7)

Airbyte has an official Redshift destination using S3 staging and the COPY command for high-throughput bulk loads — avoids row-by-row INSERT overhead.

The dbt-redshift adapter compiles dbt models into Redshift SQL and builds them as tables and views directly in the warehouse.

Fivetran loads raw data from SaaS apps and databases into Redshift and keeps it in sync incrementally, ready for modelling in the warehouse.

Airflow orchestrates Redshift queries, cluster operations, and S3-to-Redshift loads through operators in the apache-airflow-providers-amazon package.

Redshift is QuickSight's primary analytical data source, accessed via SPICE import or live Direct Query.

Excel connects to Amazon Redshift through the Redshift ODBC driver and Power Query, pulling query results into worksheets or the data model for refreshable reports.

Required by Amazon Redshift(1)

Amazon Redshift is a fully-managed AWS data warehouse — it only runs inside an AWS account and depends on AWS-native infrastructure (S3 for backups/Spectrum, IAM for access control, VPC networking); there is no way to provision Redshift outside AWS.

Alternatives to Amazon Redshift(5)

BigQuery (Google) vs Redshift (AWS) — same cloud DWH category; cloud-vendor affiliation is often the deciding factor.

Snowflake is multi-cloud with separated storage/compute; Redshift is AWS-native — both are major cloud DWH platforms.

ClickHouse can replace Redshift when AWS lock-in is a concern or when sub-second query latency on large datasets is required; Redshift's S3 and Glue integrations have no direct equivalent in ClickHouse.

Amazon Redshift is a cloud OLAP data warehouse; PostgreSQL is a general-purpose OLTP database. PostgreSQL is a practical alternative for teams whose analytical query volume does not justify a dedicated warehouse.

Databricks overlaps with Redshift on analytics; Databricks adds ML/Spark workloads, Redshift is a focused AWS-native DWH.

Vendor

Amazon Web Services

Amazon Web Services

Website →

Tags

SQLData EngineeringData Pipelines

Details

Maintained
Yes
DB model
Relational
Query language
SQL
Hosting
Cloud managed
ACID compliant
Yes
Replication
Yes