From 618ab4014798c99f36cbef24764216688a427736 Mon Sep 17 00:00:00 2001 From: counterweight Date: Thu, 23 Jul 2026 13:30:16 +0200 Subject: [PATCH] interesting article --- interesting/postgres-survival-tips.md | 14 ++++++++++++++ 1 file changed, 14 insertions(+) create mode 100644 interesting/postgres-survival-tips.md diff --git a/interesting/postgres-survival-tips.md b/interesting/postgres-survival-tips.md new file mode 100644 index 0000000..322febc --- /dev/null +++ b/interesting/postgres-survival-tips.md @@ -0,0 +1,14 @@ +# Postgres survival tips + +A guide built by the ratchet CEO with some tips from his own experience on building on top of Postgres: https://hatchet.run/blog/postgres-survival-guide. + +Most stuff I already knew, some I found interesting, other bits I disagree. + +- Stuff I liked + - "Use foreign keys with cascading deletes for low-volume tables, particularly where database consistency and correctness are important. Careful at higher volume." -> Very good call on the danger with the volume. People think foreign keys come with no trade-offs because they haven't dealt with a huge table that has them. + - Paying attention to compound indices + - Thi reference to expand and contract idea (also known as parallel change, but I prefer the former). Powerful. I had an intuition for this but never put a name on it: https://martinfowler.com/bliki/ParallelChange.html + - He also uses Dalibo's query plan visualizer!!! + +- Stuff I would challenge + - Using an incremental id on tables. This introduces a very limiting factor where data lacks unique identification outside of your database context, and distributed usage becomes hard.