One database for everything: the case against Redis, Mongo, and Pinecone
A new site catalogs a Postgres replacement for nearly every specialized data system, from Redis to Neo4j, and calls reaching for them premature optimization.
The catalog is specific: UNLOGGED tables instead of Redis, SKIP LOCKED instead of a job queue, pgvector instead of Pinecone, PostGIS instead of your geo stack, with a rule that you may leave Postgres only after documenting why it wasn't enough.
A developer running an email CRM on Redis isn't buying the premise: it's trivial to set up and never crashes, so what difficulty is this genre even describing?
Postgres for the cache, the queue, the vector store, the graph
The argument is blunt: most teams reach for a second database, a search cluster, and a vector store before the first one is even warm, and it's all premature optimization. Push PostgreSQL to its limits first, then pay for something specialized only if it truly can't keep up. The site reads as an explicit descendant of Choose Boring Technology (2018).
What makes it land is the replacement map. For almost every specialized system, the site names a Postgres feature or extension that covers the same ground:
| Specialized tool | Postgres replacement |
|---|---|
| Redis (caching) | UNLOGGED tables, materialized views |
| Job queue | SKIP LOCKED, pgmq, pgflow |
| Elasticsearch / Algolia | tsvector, pg_trgm, ParadeDB |
| MongoDB | JSONB, FerretDB |
| Pinecone / Weaviate | pgvector, pgvectorscale |
| Time-series | TimescaleDB, pg_partman |
| Snowflake / BigQuery | pg_analytics, DuckDB |
| Neo4j | Apache AGE, recursive CTEs |
| Geospatial | PostGIS |
The bar for ever leaving is deliberately high: push Postgres to its limits, document why it was insufficient, and accept the operational cost of whatever replaces it.
The people who actually run the other stuff pushed back
A developer building an email-driven CRM at carom.io doesn't buy the premise:
But I read the (growing) genre of 'use Postgres for everything' articles and they imply a difficulty in running other software that I just don't see. I'm thinking of Redis in particular… it's trivial to set up and it ~never crashes or requires maintenance.
One commenter wasn't impressed by the slogan itself:
I've seen a few 'Use Postgres for Everything!' posts lately… when I see things like 'All you need is X' I roll my eyes a little, because [the ideas of don't-add-dependencies and prefer-simplicity have] very little to do with postgres.
A reader citing an Oxide and Friends episode and brandur's writeup on Postgres queues flagged the venue:
Given the unanimous consensus on HN that Postgres is all you will ever need… Online forums like Hacker news can be a bit echo chamber-y.
One more voice took the article's own backup suggestion and turned it around:
The reason for separate systems is that some of them go down. And by 'go down', that includes you wanting to deploy a new version. Or restore from backup as the article suggested.
The uncomfortable corollary: one database means one blast radius.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
The EU is forcing Nintendo to make Switch 2 batteries user-swappable
The compliance headline has a quieter twin, buried in the same support FAQ: after nearly a decade on sale, Nintendo is retiring the original Switch, Switch Lite and Switch OLED.