Amazon Redshift
7.3/10Amazon Redshift is a managed, columnar cloud data warehouse on AWS for running SQL analytics over warehouse and data lake data at petabyte scale.
Data last updated November 2026 · Our methodology
Overview
Amazon Redshift is the data warehouse service in Amazon Web Services. It runs SQL analytics on columnar storage spread across many nodes, and it can also query data lake files in S3, including Apache Iceberg and Parquet tables. AWS first offered it in late 2012 and made it generally available in February 2013. It is based on technology from ParAccel and an older version of PostgreSQL. You can run it in two ways: as provisioned clusters that you size, resize, schedule, or pause, or as Serverless, which scales compute automatically and stops billing when idle. Pricing is usage based. Serverless costs $0.375 per RPU-hour, with a minimum of about $1.50 per hour. Provisioned on-demand clusters start at $0.543 per hour, depending on node type. Managed storage costs $0.024 per GB-month, and Spectrum lake scans cost $5 per TB. Reserved Instances and Serverless Reservations cut costs for steady workloads. Zero-ETL integrations bring in data from Aurora, RDS, and DynamoDB, and streaming ingestion works with Kinesis and MSK. Redshift ML trains SageMaker models from SQL, and data sharing works across warehouses. It connects to BI and ETL tools such as QuickSight, Tableau, Power BI, Looker, dbt, and Informatica. It fits best for companies whose data platform already runs on AWS. Its main limitations are bills that can be hard to predict, the work of tuning clusters and keys such as DISTKEY and SORTKEY, and lock-in to AWS-specific features, which makes later migration harder.
Our Rating
Who Amazon Redshift is best for
Best software for EnterpriseGood fit
It is built for petabyte-scale MPP warehousing, with VPC isolation, SSO, and row- and column-level security that large organisations need.
It is a managed service, but teams still have to size clusters, tune it, and manage IAM access.
It suits SaaS companies on AWS that need zero-ETL analytics on their Aurora, RDS, or DynamoDB data.
Less suited
Serverless and the $300 trial credit lower the cost of starting, but bills can be hard to predict for small teams.
Pros & Cons
Pros
- Zero-ETL from Aurora, RDS, and DynamoDB means AWS teams can analyse operational data without building pipelines.
- There are two deployment modes: Serverless shuts down when idle, while provisioned clusters can be paused or scheduled.
- Reviewers on PeerSpot praise how it handles very large datasets and scales up or down.
- Security controls include VPC isolation, AES-256 encryption, SSO, row- and column-level access, and data masking.
- Multi-AZ deployments carry a 99.99% availability SLA.
Cons
- Spectrum scans and concurrency scaling beyond the free credits are billed per use, which makes costs hard to predict.
- Teams must size clusters and tune DISTKEY and SORTKEY, and reviewers describe setup as complex.
- PeerSpot reviewers report slowdowns with dynamically generated or parallel queries, and instability on large clusters.
- Redshift-specific features make migrating away from it, or adopting multi-cloud, harder.
- Hudi and Delta Lake tables are read-only from Redshift.
Key Features
Amazon Redshift Pricing
Free trial (Serverless)
- Credit expires after 90 days
- $300 credit for first-time Serverless users
- Two-month provisioned trial capped at 750 hours per month in regions without Serverless
Provisioned (on-demand)
- Starts at $0.543 per hour
- ra3.4xlarge costs $3.26 per node-hour
- Clusters can be resized, paused, or scheduled
- Reserved Instances available on 1- or 3-year terms
- One hour of free concurrency scaling credit per day
Serverless
- $0.375 per RPU-hour
- Starts at about $1.50 per hour
- Base capacity from 4 to 1024 RPUs
- Billed per second with a 60-second minimum
- Reservations save up to 24% on a 1-year term or 45% on a 3-year term
Frequently Asked Questions
Yes. First-time Serverless users get a $300 credit that expires after 90 days. In regions without Serverless, there is a two-month provisioned trial capped at 750 hours per month.
Serverless costs $0.375 per RPU-hour, billed per second with a 60-second minimum. Reservations cut the cost by up to 24% on a 1-year term or 45% on a 3-year term.
No. Managed storage is billed separately at $0.024 per GB-month, and Spectrum lake queries cost $5 per TB scanned.
Multi-AZ data warehouses carry a 99.99% availability SLA.
Yes. The Redshift Data API runs SQL asynchronously through a secured endpoint and keeps results for 24 hours.
It works with QuickSight, Tableau, Power BI, Looker, Alteryx, dbt, and Informatica, among others.
Similar Tools
See all alternatives
Amazon's public cloud platform, with over 200 pay-as-you-go services for compute, storage, databases, networking and AI, used by startups through large enterprises and governments.
Northflank is a Kubernetes-based developer platform for deploying services, jobs, databases and GPU workloads on its managed cloud or on customers' own cloud accounts.
AWS Amplify is a pay-as-you-go platform from Amazon Web Services for hosting web apps from Git and building full-stack backends on AWS.
Ranked near Amazon Redshift in Hosting & Infrastructure
See Amazon Redshift in the listAmazon Redshift is #19 in Hosting & Infrastructure by our score.
