February 23, 2026
Who let the DB dogs out?
Show HN: PgDog – Scale Postgres without changing the app
PgDog promises painless Postgres scaling, fans cheer while skeptics bark about lag
TLDR: PgDog is a Rust tool that helps apps scale Postgres by pooling connections, balancing traffic, and sharding data with one endpoint. Commenters praised it for pooling but debated latency and read/write safety under replication lag, asked about two‑phase commits, and wondered if fintech is the main audience.
PgDog just strolled into Hacker News saying it can help apps scale the popular Postgres database without changing a single line of code. It’s a Rust-built “traffic cop” that juggles connections, spreads the load, and even shards your data into pieces. The crowd? Equal parts tail-wagging and side-eye. One happy user swore by it as a connection pooler (the tool that lets many app users share a few database connections), and hinted sharding tests look promising—but not battle‑tested yet. Another cheerleader wanted it on the “Postgres Is Enough” list, a site celebrating teams that go all‑in on Postgres.
Then the drama unleashed. A fan loved PgDog’s clever cross‑shard math—like rewriting averages so they work across split data—but asked about latency: how much delay does the dog add when you just want a simple read? Another commenter poked the biggest bear: read/write splitting sounds great until replication lag (copies of data arriving late) bites. Do you need the strict, slower “synchronous” mode to be safe? And someone clocked Coinbase and Ramp logos on the homepage, pressing whether fintechs are the prime targets. Sprinkle in dog puns and cautious vibes, and you’ve got a crowd that’s excited—but demanding receipts before this pup guards production.
Key Points
- •PgDog is a Rust-based proxy that scales PostgreSQL via connection pooling, load balancing, and full database sharding.
- •It offers deployments via Kubernetes (Helm), AWS (EKS/ECS with Terraform), and Docker, with a quick-start demo including 3 shards and 2 sharded tables.
- •Configuration uses pgdog.toml for database hosts and sharding, and users.toml for credentials; pools are created only when users exist for a database.
- •PgDog supports transaction and session pooling, handling SET statements and startup options, and provides recovery features like auto rollback and re-synchronization.
- •As an OSI Layer 7 load balancer, PgDog routes writes to the primary and reads to replicas using query parsing, with health checks removing unhealthy hosts from rotation.