📣 We have a brand new Postgres platform with: Instant Copy-on-Write branching Built-in data anonymization Separation of storage and compute 100% vanilla Postgres It’s for staging/dev environments as well as for production workloads. Details in 🧵

A week in Rome with the Xata crew 🇮🇹 Less meetings, more real talk, better ideas. We turned a Roman villa into HQ: pasta by night, plans by day. Rome gave us time to connect, think, and play, then ship faster. What makes an offsite great? Team chemistry > calendar invites. ✨

Day 1 at #PGConfEU kicked off with Community Summits! Our own Gülçin Yıldırım Jelínek is co-organizing two of them ☸️ PostgreSQL on Kubernetes Summit 🧠 PostgreSQL & AI Summit Big thanks to the volunteers & speakers who made this happen. Community is what makes Postgres great

Next week we’re heading to #PGConfEU, the biggest Postgres event in the world. On Oct 21, our own Gulcin Yildirim Jelinek co-running two community summits: 🧠 PostgreSQL & AI Summit ☸️ PostgreSQL on Kubernetes Summit Then on the main track, catch Esther Minano & Ahmet

Offsite and smiles, my favorite time of the year with the @xata team

Branch settings, faster. We reorganized the page and added search so you can find the setting you need without scrolling. What changed Search to jump directly to a setting. Logical groups with clear labels. Guardrails to reduce misclicks. Why it helps Quicker navigation. Less

See write latency and replica lag before users do. New performance charts surface: WAL sync time → time to flush WAL to disk (disk/flush latency). Replication lag → how far a standby is behind the primary (freshness). Why it matters Early warning for slow writes (spikes in

Create a Postgres branch in one screen. Name it → get the branch → run queries. No org/project setup first. Sane defaults now, easy rollback later. Want early access? Ping us and we’ll enable it.

Recap/slides/repos 👇 https://t.co/FqKxtwyqZA

Getting paged at 3 a.m + database operations ≠ heroics. At the Berlin Postgres meetup we showed Xata Agent to: • Autonomously find slow queries • propose safe fixes (indexes/config/query optimization) • run on a schedule w/ approvals It is like having a database expert

Preload extensions once, use them everywhere. Set shared_preload_libraries at the branch level and every database on that branch shares the same foundation. Example: pg_stat_statements, pg_trgm → consistent query stats + text ops across envs. Notes: applies to all DBs on the

Project-level default database now live in Xata. Set one default and we point new connection strings, views, and child branches to it by default. Fewer misroutes and accidental writes to secondary DBs; faster onboarding. Set it in Project Settings → Default database. You can

Full blog post with examples for each of the above:

I wrote a long post going in detail through ~30 PostgreSQL 18 new features and changes. Beyond the highlight features (async io, oauth, uuid, etc.) there are tons of small goodies. Here are some of my favorites that you might not have heard about yet.

PostgreSQL 18 is supported on Xata. Try PG18 in a safe, branch-based workflow: How to start: • Clone with anonymization → test on look-alike data. • Create a PG18 branch → run queries and migrations. • When ready, promote with pgroll for zero-downtime changes. Why test

Tudor Golubenco: Going down the rabbit hole of Postgres 18 features

Wrote a short post about how we built the Text to SQL feature at Xata

Postgres 18 AIO config settings on @xata.

Not every dev branch needs to be alive 24/7. At @xata , we now hibernate them automatically after inactivity (default 30 min, configurable). Only pay when you actually use them.

Full notes: