Honker is a SQLite extension that adds message broker functionality—including queues, streams, pub/sub, and cron scheduling—directly into SQLite database files. The project gained 170 points and 50 comments on Hacker News on April 30, 2026.
Sub-Millisecond Wake Latency Using SQLite's Data Version Counter
Honker monitors SQLite's PRAGMA data_version counter, a monotonic value that SQLite increments on every commit from any connection. By polling this counter approximately 1,000 times per second, Honker achieves cross-process wake latency of roughly 0.7 milliseconds at the median on M-series laptops.
This approach enables event-driven messaging without external infrastructure while maintaining SQLite's ACID guarantees. Idle operational cost involves one lightweight SELECT query per millisecond per database with minimal page-cache pressure.
Atomic Transactions Eliminate Dual-Write Consistency Problems
The extension's key technical advantage is seamless integration with business logic through atomic transactions. When developers enqueue a job within a transaction containing an INSERT statement, both operations commit together or both rollback. This eliminates dual-write consistency problems between application data and queue state that plague systems using separate databases and message brokers.
Honker provides task queuing with job visibility and retry logic, event streams for temporal data, PostgreSQL-style NOTIFY/LISTEN pub/sub semantics, and scheduled job execution with cron-like functionality—all with zero polling overhead for idle listeners.
Cross-Language Support With Unified On-Disk Format
The extension works across Python, Node.js, Rust, Go, Ruby, Bun, and Elixir while maintaining a unified on-disk format. This enables polyglot applications to share queue and pub/sub infrastructure through a single SQLite file.
Listener scaling is inherently efficient since wake signals distribute through a single shared polling mechanism rather than individual queries per listener. This design makes Honker particularly suitable for applications already using SQLite for primary data storage that need job processing without introducing separate infrastructure.
Targets Edge Deployments and Infrastructure Simplification
The project references real-world SQLite deployments including Bluesky's PDS, Fly's LiteFS, and Turso as examples of systems that could benefit from embedded message broker capabilities. Unlike Redis or Celery, Honker requires no separate infrastructure to maintain—everything lives inside the existing SQLite file.
Hacker News discussion focused on use cases for edge computing, embedded systems, and reducing infrastructure complexity. Developers noted that Honker fills a gap for applications that don't need Redis-scale throughput but want reliable queuing with ACID guarantees.
Key Takeaways
- Honker embeds durable queues, streams, pub/sub, and cron scheduling directly into SQLite files, gaining 170 points on Hacker News in April 2026
- The extension achieves 0.7ms median wake latency by monitoring SQLite's data version counter approximately 1,000 times per second
- Atomic transactions integrate queue operations with business logic, eliminating dual-write consistency problems between application data and queue state
- Cross-language support includes Python, Node.js, Rust, Go, Ruby, Bun, and Elixir with a unified on-disk format
- Ideal for applications already using SQLite that need message broker capabilities without deploying separate infrastructure like Redis or RabbitMQ