Skip to main content

Databricks Partner Well Architected Framework

The Partner Well-Architected Framework provides comprehensive architectural guidance for technology partners building solutions with the Databricks Data Intelligence Platform. This framework defines best practices, design patterns, and validation standards for integrations, data products, and applications built on Databricks.

Successful technology partnerships are built on a foundation of technical excellence. Well-architected systems and high-quality integrations drive mutual business success and create lasting value for our joint customers.

Who This Guide Is For​

This framework is for architects, developers, and product leaders at technology partner organizations.

You should use this guide if you are:

Connected ISV Partners​

Building integrations that connect your product to Databricks—such as BI tools, data ingestion platforms, governance solutions, or ML platforms.

See Connected ISV Partners

Data Collaboration Partners​

Sharing data privately or through the Databricks Marketplace using Delta Sharing to monetize and distribute data products.

See Data Collaboration Partners

Built-On ISV Partners​

Building your core product on Databricks as the intelligence layer—including SaaS platforms, embedded analytics, and AI-powered applications.

See Built-On ISV Partners

Building on Established Frameworks​

The Partner Well-Architected Framework builds upon Cloud Well-Architected Frameworks (AWS, Azure, GCP) and the Databricks Lakehouse Architecture, extending these established practices to address partner-specific challenges: multi-tenant SaaS architectures, customer data integration, marketplace distribution, and partner attribution.

Pillars of the PWAF​

The Partner Well-Architected Framework is built on four core architectural pillars:

1. Architectural Best Practices — This guide provides design principles, repeatable design patterns, reference architectures, and implementation guidance for building solutions with Databricks.

2. Defined Technical Standards — Each partner type has specific technical standards that must be met for validation and program participation, along with a clear path to validation.

3. Measurement & Attribution — Continuous measurement of adoption, impact tracking, and partner attribution through telemetry to understand usage patterns and joint business impact.

4. Operations & Lifecycle Management — Operational guidance, monitoring, automation, and lifecycle best practices for maintaining partner solutions over time, including runbooks, access management, and continuous improvement workflows.

Built for Modern AI Tooling​

This documentation is optimized for AI assistants and modern development workflows, enabling you to quickly find answers and implement solutions using your preferred AI-powered tools. This framework feature applies across all architectural pillars, making the entire framework accessible and actionable through AI-powered development workflows.

Learn more about AI-Ready Documentation

Build with Firefly​

Firefly Analytics is a production-ready reference implementation showcasing PWAF patterns in action. Firefly is an open-source SaaS application built entirely on Databricks with custom authentication, multi-tenancy, and embedded apps.

Throughout the Built-On sections, you'll find references to Firefly with live code examples you can use directly or point your AI tools at to accelerate your development.

Learn more about Firefly Analytics

Join the Partner Program​

This guide is designed for technology partners building with Databricks. To receive go-to-market benefits, co-sell opportunities, and program incentives, you must first be an active member of the Databricks Partner Program.

Visit the Databricks Partner Portal to apply, access program resources, and submit validation requests. Need architectural guidance or have questions about your integration? The Partner Engineering team is here to help—submit a Partner Support request through the Partner Portal to connect with our technical experts.

Get Started​

Explore the main sections of this framework: