multigres: Scale Postgres with a Vitess-Inspired System

Overview

Multigres is an early-stage project adapting Vitess for PostgreSQL. It aims to help teams manage PostgreSQL deployments that need to scale beyond a single database instance, bringing a Vitess-inspired approach to PostgreSQL workloads.

Its stated areas include connection pooling, sharding, horizontal scaling, and high availability. It may interest teams planning for those needs, but its early maturity makes it more suitable for evaluation than for assuming production readiness.

Key Features

• Adapts the Vitess project’s approach for PostgreSQL.
• Targets connection pooling for PostgreSQL deployments.
• Focuses on horizontal scaling and sharding.
• Addresses high-availability database setups.
• Provides an EKS getting-started guide for running Multigres on Kubernetes.
• Publishes PostgreSQL regression, isolation, and extension test status through its project badges.

Use Cases

• PostgreSQL operators evaluating a Vitess-inspired path to horizontal scaling.
• Teams investigating sharding as their PostgreSQL workload grows.
• Platform engineers exploring high-availability PostgreSQL architectures.
• Kubernetes teams wanting to evaluate Multigres on EKS using the project guide.

Project Facts

• Language: Go
• License: Apache-2.0
• Stars: 2.7k
• Forks: 146
• Topics: connection-pooling, database, distributed-systems, high-availability, horizontal-scaling, postgres, postgresql, sharding, vitess
• Archived: No

Getting Started

Visit the Multigres site to try the project. For an EKS setup, follow the EKS getting-started guide. See the repository README for project details.

Alternatives

• vitess: Vitess scales MySQL through sharding, while Multigres applies Vitess-style database management to PostgreSQL.

Considerations

The project describes itself as being in the early stages of development. Its contribution policy currently welcomes minor bug fixes and feedback, while major contributions are not yet open. Review its maturity and fit for your needs before relying on it for production workloads.