1/11

13 Nov 2026

QuickSight Data Sources: Query Live or Import to SPICE?

Amazon QuickSight connects to AWS services such as Athena, Redshift, Aurora, RDS, and S3; to databases and warehouses outside AWS such as Snowflake, Teradata, Databricks, and Google BigQuery; to uploaded or S3-hosted files in CSV, TSV, ELF, CLF, JSON, and XLSX formats; and to a short list of SaaS sources including Salesforce, ServiceNow, and Adobe Analytics. For each source, you can either query it live or import it into SPICE, QuickSight’s in-memory engine. AWS updates the supported list often, so confirm the current list for your region and edition before you design around a connector.

Most teams start a QuickSight project by asking whether it can connect to their data. A better question is how it should connect. The same source can be queried live, which keeps data fresh and in place, or imported into SPICE, which makes dashboards fast and takes load off the source. That choice shapes speed, cost, and governance far more than the connector list does.

This page explains which data sources QuickSight supports, what changed with the QuickSight name, how to choose between live queries and SPICE, and what governed and public sector teams should check. If you are comparing platforms, see our QuickSight vs Tableau comparison, and if you are moving from Tableau, our Tableau to Amazon QuickSight migration article covers the mapping.

Quick Summary

QuickSight supports AWS services, third-party databases and warehouses, files, and a few SaaS sources. Each source connects either live or through SPICE, and that decision drives performance, cost, and freshness. Check AWS’s current list, plan security and networking early, and test your largest dataset before you commit.

data-sources-supported-in-amazon-quicksight

Key Takeaways

The list is broad but changes. AWS adds and retires connectors, so check the current documentation, not an old blog post.
Live or SPICE is the real decision. It sets freshness, speed, and cost for every source.
Not everything must be in AWS. QuickSight reaches Snowflake, Teradata, Databricks, BigQuery, and on-premises databases, though some AWS sources must live in AWS.
Security starts at the connection. Plan IAM, private networking, row-level security, and encryption before you build dashboards.
Test your largest dataset. A proof of concept on real volumes beats any support list.

QuickSight Data Sources at a Glance

Category	Examples	Typical Way to Connect
AWS services	Athena, Redshift, Redshift Spectrum, Aurora, RDS, S3, S3 Tables, OpenSearch	Direct query or SPICE
Warehouses and lakehouses	Snowflake, Databricks, Google BigQuery, Teradata	Direct query or SPICE
Relational databases	SQL Server, Oracle, PostgreSQL, MySQL, MariaDB	Direct query or SPICE, often over private networking
Files	CSV, TSV, ELF, CLF, JSON, XLSX	Upload, or read from S3 with a manifest
SaaS applications	Salesforce, ServiceNow, Adobe Analytics	Import, usually on a refresh schedule

What this means for you: Find your sources in the table, then decide for each one whether freshness or speed matters more.

A Note on Names: QuickSight and Quick Suite

AWS has folded QuickSight into Amazon QuickSight, with new branding rolling out from October 9, 2025, and AWS documentation now labels the service “Amazon QuickSight.” The BI engine and SPICE are unchanged. This page uses “QuickSight” because it is still the name most people search for, and AWS notes that some newer capabilities are available only in certain regions and editions.

AWS Services: The Native Path

QuickSight connects most easily to data already in AWS:

Amazon Athena queries data in S3 with SQL, so a data lake on S3 becomes a QuickSight source in minutes. Athena also reads the AWS Glue Data Catalog.
Amazon Redshift and Redshift Spectrum serve warehouse-scale analytics, and Spectrum extends queries into S3.
Amazon Aurora and RDS let teams connect to existing relational databases without building extraction pipelines.
Amazon S3 works through a manifest file that tells QuickSight where files live and how to read them, and AWS now lists S3 Tables as a source as well.
Amazon OpenSearch Service adds search and log analytics data.

AWS documents that Redshift clusters, Athena databases, and RDS instances must be in AWS. A well-organized data lake on S3 usually simplifies this path, and our AWS consulting services and Redshift consulting services pages describe how we build it. Teams that want to query Spark can also read our article on Apache Spark as a data source in QuickSight.

Databases and Warehouses Outside AWS

QuickSight is not limited to AWS. AWS’s documentation and FAQs list Snowflake, Teradata, Databricks, and Google BigQuery, and QuickSight also connects to relational databases such as SQL Server, Oracle, PostgreSQL, MySQL, and MariaDB. That lets organizations keep their current architecture while adding QuickSight dashboards on top.

A few cautions apply. Databricks support is documented for specific platform and Spark versions, so check yours. On-premises or private databases usually need private connectivity, such as a VPC connection. If your warehouse is Snowflake or Databricks, our Snowflake consulting and Databricks migration services pages show how those platforms feed BI.

What this means for you: Check version and networking requirements for each non-AWS source early, because they are the usual cause of delays.

Files and SaaS Sources

Files. QuickSight reads CSV and TSV text files, ELF and CLF log files, JSON (flat or semi-structured), and XLSX workbooks. You can upload them directly or keep them in S3 and point QuickSight at them with a manifest. QuickSight infers schema and types from JSON, and AWS notes it supports UTF-8 encoding but not UTF-8 with a byte order mark.

SaaS applications. AWS’s current documentation lists Salesforce, ServiceNow, and Adobe Analytics as SaaS sources. Older guides also name connectors such as Twitter, so check AWS’s current list before you plan around one. SaaS data is usually imported and refreshed on a schedule, which makes it near-current but not live.

Query Live or Import to SPICE?

Every dataset that supports both modes forces the same choice:

Direct query keeps data in the source, so dashboards show the latest values. It depends on the source’s speed and adds query load, so slow or busy sources can slow dashboards.
SPICE imports the data into QuickSight’s in-memory engine, so dashboards respond quickly, and the source is protected from heavy query traffic. Data is only as fresh as the last refresh, and SPICE capacity beyond what is included is billed separately.

A sensible rule is to use SPICE for most dashboards that many people view, and direct query where freshness matters more than speed or where data must stay in place. Our article on unique features of Amazon QuickSight covers SPICE and related features, and the Standard vs Enterprise editions article explains which capabilities depend on edition.

What this means for you: Decide live or SPICE per dataset, based on how often it changes and how many people view it.

Security and Governance for Connected Data

Connecting many sources raises the stakes on security. QuickSight works with IAM permissions, VPC connections, AWS KMS encryption, and row-level security, which restricts what each user sees based on their role. Row-level security is part of the Enterprise edition. Our encryption of data in QuickSight article covers protection at rest and in transit, and our guides to migrating row-level security and data security compliance explain how to plan the rules.

Public Sector and Regulated Considerations

This section is planning guidance, not a statement of DataTerrain’s authorization status or government work. Public sector and other regulated teams should check four things before connecting a source:

Region and authorization. Confirm that the QuickSight edition, the region, and each connected service fit your authorization boundary.
Data residency. Keep data in approved regions, and check whether SPICE copies are stored where your rules require.
Private connectivity. Prefer private networking over public endpoints for sensitive sources.
Auditability. Keep logs of who accessed what, and keep dashboard definitions as records where retention rules require.

How to Plan Your QuickSight Data Architecture

A short, practical sequence avoids most surprises:

Inventory sources and consumers, including owners, volumes, and refresh needs.
Choose a path per source: native AWS, warehouse connection, file, or SaaS import.
Decide live or SPICE for each dataset.
Prepare the data upstream where you can, since heavy joins and cleaning belong in a pipeline, not in every dashboard. Tuning the pipelines that feed BI pays off here.
Set security, then validate. Apply row-level rules, test with real roles, and reconcile key figures against the source. Our BI reports and dashboard development service supports this stage.

Frequently Asked Questions

Which data sources does Amazon QuickSight support?

QuickSight supports AWS services such as Athena, Redshift, Aurora, RDS, and S3; databases and warehouses such as Snowflake, Teradata, Databricks, Google BigQuery, SQL Server, Oracle, PostgreSQL, MySQL, and MariaDB; files in CSV, TSV, ELF, CLF, JSON, and XLSX formats; and SaaS sources such as Salesforce, ServiceNow, and Adobe Analytics. AWS updates the list, so confirm it for your region.

Does QuickSight connect to Snowflake, Databricks, and Google BigQuery?

Yes. AWS’s documentation and FAQs list Snowflake, Databricks, and Google BigQuery as supported sources, along with Teradata. Databricks support is documented for specific platform and Spark versions, so check your version and networking requirements.

What is the difference between direct query and SPICE?

Direct query reads data from the source each time, so it is fresh but depends on the source’s speed. SPICE imports data into QuickSight’s in-memory engine, so dashboards are fast, and the source is protected, but data is only as current as the last refresh.

Can QuickSight connect to on-premises databases?

Yes. AWS lists on-premises databases such as SQL Server, MySQL, and PostgreSQL as supported, usually over private connectivity such as a VPC connection. Check networking, authentication, and version requirements with your security team.

What file formats can QuickSight import?

QuickSight imports CSV and TSV files, ELF and CLF log files, JSON files, and XLSX workbooks, from a local upload or from S3 using a manifest file. It supports UTF-8 encoding but not UTF-8 with a byte order mark.

Scope Your QuickSight Project with DataTerrain

The costly mistake is connecting everything first and deciding about speed, cost, and security afterward. DataTerrain brings 17+ years and 400+ clients in BI and reporting modernization, with QuickSight deployment, data source design, security configuration, and validation of key figures against the sources you already trust.

Talk to our BI team