~postgresweekly | Issue 584 - Don't do this in Postgres (but do open this newsletter) (8)

Don't Do This (in Postgres) — An interesting page on the official Postgres wiki that rounds up a variety of advice around ‘common mistakes’ in using Postgres and what not to do, such as “Don’t use char(n)” and “Don’t use serial“. Some of these will be contentious, but are well backed up with reasons. The page is updated, but it’d be nice to see...

The Power of Open Source in Postgres — Laurenz demonstrates how Postgres’ open source nature is invaluable when it comes to answering questions about how it operates, in case by tracking down the reason for the default setting for autovacuum_vacuum_scale_factor. Knowing how to do what he does in this post is a valuable skill.

Find the Best Postgres Monitoring & Tuning Vendor for 2025 — Download our free Vendor Evaluation Template to swiftly compare features, score demos, and confidently pick the best Postgres performance & monitoring solution to meet your 2025 budget needs.

The Anatomy of Table-Level Locks — A look at the meaning and motivation behind locking in database systems, before more closely looking at the various ways tables can be locked in Postgres and how conflicts between these locks are resolved.

Idle Transactions Cause Table Bloat? Wait, What? — “Yup, you read it right. Idle transactions can cause massive table bloat that the vacuum process may not be able to address.”

pg_mooncake: Fast Columnstore Tables in Postgres for Analytics — A Postgres extension that adds columnar storage and vectorized execution (using DuckDB) for fast analytics within Postgres. Columnstore tables are stored as Iceberg or Delta Lake tables on the local file system or cloud storage. v0.1 just landed.

🐕 pgDog: A New Rust-Powered Postgres Query Router and Proxy — The main developer of the PgCat Postgres pooler has shifted to working on this dog-themed ‘spiritual successor’ that offers similar pooling, load balancing, and proxying, but with a plugin system and an AGPL license (PgCat is MIT).

Ruby Active Record Adapter for AWS Aurora DSQL — The “very beginnings” of an Active Record (Ruby on Rails) connection adapter for Amazon’s new Aurora DSQL database (essentially a Postgres-compatible serverless, distributed database service).