Liftbridge

Liftbridge is an open-source distributed message streaming system (messaging/streaming) that provides durable, ordered, log-based streams built on top of NATS.

• Distributed message streaming built on NATS with durable, append-only logs (messaging/streaming).
• Horizontally scalable streams with replication for fault tolerance (high-availability infrastructure).
• At-least-once delivery semantics and message retention policies (data durability and reliability).
• Stream partitioning via subjects and stream shards for parallelism (scalable data pipelines).
• gRPC-based APIs and NATS integration for client access and ecosystem interoperability (integration middleware).

More About Liftbridge

Liftbridge is an open-source distributed message streaming system (messaging/streaming) designed to provide durable, ordered logs on top of NATS, which is a lightweight, cloud-native messaging system. It targets use cases where applications require durable event streams, log-based messaging, and replayable histories while retaining the operational model and protocol of NATS. By layering log semantics over NATS subjects, Liftbridge enables event sourcing, streaming Extract, Transform, Load (ETL), and other log-centric patterns without replacing an existing NATS deployment.

The core capability of Liftbridge is the provision of named streams (log-based messaging) that are attached to NATS subjects and backed by append-only logs. Each stream can be configured with replication factors, retention policies, and partitioning. Streams support message ordering within partitions and at-least-once delivery semantics (reliable messaging). Liftbridge uses a Raft-based metadata store (distributed coordination) to manage stream configurations, leadership, and cluster membership, which provides consistency for control-plane operations.

Liftbridge supports horizontal scalability by allowing streams to be sharded into multiple partitions (scalable data pipelines). Each partition is replicated across brokers, with a designated leader responsible for handling writes and serving reads, while followers replicate data for redundancy. This model supports failover at the partition level, helping maintain stream availability during broker outages. Message retention policies can be configured based on time or size, controlling how long messages remain in the log for replay or backfill processing.

Client interaction with Liftbridge occurs primarily through gRPC APIs (service-to-service communication) and through integration with NATS clients (messaging integration). Producers can publish messages directly to Liftbridge streams or indirectly via NATS subjects that Liftbridge is configured to consume and log. Consumers can read from specific offsets, start at the earliest or latest position, and use cursor-like subscriptions for continuous streaming consumption. This enables event replay, catch-up processing, and coordinated consumption patterns in distributed systems.

In enterprise environments, Liftbridge can serve as a durable streaming layer for microservices, analytics pipelines, and event-driven architectures (enterprise integration). Organizations using NATS for low-latency messaging can introduce Liftbridge when durable storage, ordered logs, and replay are required, without migrating to a different protocol. The project’s positioning is as a log-based streaming system, integrating into categories such as message streaming platforms, event log infrastructure, and distributed log storage. Its interoperability with NATS and use of gRPC allow it to participate in polyglot, multi-language environments and to integrate with container-orchestrated or cloud-native deployments.