Comment by olucasandrade
3 days ago
You've heard the advice: "Use the right tool for the right job." Sounds reasonable. Even wise. So you followed it. You picked Redis for caching, Elasticsearch for search, Kafka for messaging, MongoDB for documents, Pinecone for vectors, InfluxDB for time-series, and Postgres for... well, the relational stuff. Congratulations. You now have 7 databases to maintain, 7 backup strategies to manage, 7 monitoring dashboards to watch, 7 security audits to run, and 7 monsters that can break at 3 AM.
The thing nobody talks about, cause it doesn't sell: that advice "the right tool for the right job" is the battle cry of every vendor's marketing department.
The uncomfortable truth is that PostgreSQL is not "just a relational database." It hasn't been for over a decade. It's a data platform that does what most of these specialized tools do, using the same algorithms, with a single connection string, a single backup strategy, and a single place to debug when everything breaks at 3 AM.
Not "close enough." Not "good enough at small scale." The same algorithms.
- Redis
- Elasticsearch
- Pinecone
- Kafka
- MongoDB
- InfluxDB
All these buzzwords, unless we're talking about truly unbelievable scale, are achievable with just Postgres.
Let me show you. Actually, no... simulate it and see for yourself.
Postgres needs to figure how to solve write-amplification caused by MVCC. Any use-case with write load runs into this and then needs to spend significant time managing and working around the problem.
https://www.cs.cmu.edu/~pavlo/blog/2023/04/the-part-of-postg...
AI Slop.
Not a single human I know talks like this.
"The uncomfortable truth"? If you're uncomfortable, get checked for piles.
You're not smart. You're a lunatic.
Yeah the weird thing is you could tell the same AI to write the opposite view and it would. I don't see what the endgame is in most of these things. Ego? Street cred?
Graphs? How good is AGE really?