PostgreSQL Tooling Is Getting Serious Investment
PgDog, a Rust-written sharding, connection pooling, and load balancing layer for PostgreSQL, just announced funding. The pitch is horizontal scaling without application code changes, which is exactly what teams hitting Postgres scaling walls want to hear. The HN thread is cautiously optimistic, with the main question being what caveats come with making it 'just work' compared to standard Postgres.
This is part of a larger pattern: Postgres has become the default database for a huge swath of startups, and the ecosystem around it is maturing fast. Tools that add sharding, replication, or connection pooling without requiring a migration away from Postgres are filling a real gap. ParadeDB is also hiring in this space, and several other YC companies are building Postgres-adjacent infrastructure.
The bet these companies are making is that Postgres lock-in is real and teams will pay for tooling that extends it rather than replace it with a purpose-built distributed database.
So what?
If you're on Postgres and hitting write throughput or connection limits, the tooling options are much better now than they were two years ago. PgDog is worth evaluating before you commit to a Citus migration or a full rewrite on a distributed database. The switching cost from Postgres to anything else is high enough that extension tooling is almost always the right first move.