


The fastest and most reliable database for Postgres & Vitess Discord: pscale.link/community planetscalestatus.com


This is such a super cool new feature. Love it. I had something similar hand rolled myself for some @cursor_ai databases. This is much easier to setup, one-click to install and also tuned to work out of the box.
Your database is about to get way faster. Introducing Insights Agent. Automatically find and fix performance problems in your database and code.


Your database is about to get way faster. Introducing Insights Agent. Automatically find and fix performance problems in your database and code.

If you're not careful, a single connection can hang your entire database. Here's how it happens, how to fix it, and how to prevent it in the first place: planetscale.com/blog/debuggi…

At 150K documents with 16 searches in flight, the gap got much bigger. PlanetScale held 22 ms p95 at about 1,330 queries a second. Neon was at 233 ms p95 and about 110 queries a second through its HTTP driver, or 362 ms p95 through Hyperdrive. Giving Neon 4 CPU brought its Hyperdrive p95 down to roughly 40 to 77 ms, still above PlanetScale on 1 vCPU.


Can you complete a smooth Postgres migration without logical replication? Yes, and that's exactly what we've done for countless customers moving from Heroku to PlanetScale. How it works: planetscale.com/blog/migrate…

A well designed schema with optimized indexes is the foundation of good database performance. Combined with great hardware (local NVMe) you get an insane system. PlanetScale gives you the tools to automate figuring out the former. And the latter, well, we have it on lock.
classic, forgot an index
View quoted post ↗
PlanetScale is at KubeCon in SLC. Stop by to get a demo, grab some merch, and meet the team.

What would Postgres look like if you designed it from scratch today? And can LLMs build bbtree, the fastest B-tree around, without giving up correctness? Manish Rai Jain (Egres) and Xiangpeng Hao (SpiralDB) answer both tonight at the SF Database Meetup at PlanetScale HQ. luma.com/planet-vxwl

4 x86 vCPUs ≠ 4 ARM vCPUs Hyperthreading, vector width, clock speed, and more all impact how Postgres performs. Here's how to choose: planetscale.com/blog/when-to…


How do you ship fast while not breaking the database? Most of the principles are decades-old ideas in reliable infrastructure. The key with agents is exposing them as primitives in just the right ways. My talk from AIE this summer (YouTube link below)

epic. planetscale.com/changelog/oi…

Follow our Luma calendar to stay up to date on all our upcoming events: luma.com/planetscale