Xata 🦋

📣 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 🧵

When Postgres isn't enough for a query, pg_duckdb is the escape hatch. Runs analytical queries on the same data without an export pipeline. Read about it here: xata.io/blog/postgres-…

The trick for iterating on warehouse schema: copy-on-write branches. Test new materialized views on a branch, validate the numbers, merge to the live warehouse. Production data never at risk.

We built a product analytics warehouse in Postgres. No OLAP engine. Materialized views flatten JSONB event payloads; pg_cron refreshes them nightly.

Our CTO @tudor_g is speaking tonight at the PostgreSQL Berlin Meetup at Zalando. "From 20 seconds to sub-second: Speeding up CloudNativePG wake-ups." How we got CNPG scale-to-zero clusters waking up in under a second. Come say hi if you're in Berlin: meetup.com/postgresql-mee…

We are excited to announce that Signadot is teaming up with @xata as an integration partner! 🤝 Enabling agentic workflows at scale requires realistic environments where agents can validate changes against complex dependencies and real data.

Bullish on @xata tho!

1/3 Scaling @xata branching to your application layer just got easier. 🧵

pgstream v1.0.0 is here 🎉 A major milestone bringing stateless schema change replication to Postgres CDC. • DDL emitted directly into WAL • No schema logs • Unified snapshot handling Full details 👇 xata.io/blog/pgstream-…

Did you know Postgres lets you put constraints on domains, not just tables? A domain is a custom data type with rules attached, and Postgres stores those CHECK constraints right in pg_constraint linked by contypid instead of conrelid. A clean way to centralize data rules👌 Curious to dig deeper? Check out @apatheticmagpie’s full breakdown on the blog👉 xata.io/blog/constrain…

I’ve been waiting a long time to finally share this 😊 Today, we’re announcing that Xata has acquired @privacydynamics This is a big step for us: combining best-in-class data anonymization with Xata’s branching and developer workflows, so teams can work with production-like data without compromising on privacy or compliance. I’m incredibly excited to welcome @grahamthompson to Xata as Head of GTM, and I can’t wait to show what we’re building next 🚀

Regular triggers fire immediately and can’t be deferred. Constraint triggers can, which means Postgres can delay validation until COMMIT. The key difference is that constraint triggers integrate with the constraint system. Because of that you can control their firing time with SET CONSTRAINTS, something regular triggers don’t support. Want the full story? Read more of what @apatheticmagpie has to say in her blog 👉 xata.io/blog/constrain…

Did you know Postgres has something called a constraint trigger? Foreign keys are implemented using automatically generated constraint triggers and this is what allows them to be DEFERRABLE or INITIALLY DEFERRED. @apatheticmagpie explains the mechanics, why these triggers can only be defined as AFTER/FOR EACH ROW and why they exist in the first place. 👉 xata.io/blog/constrain…

We just shipped AI-assisted PII removal for xata clone. It analyzes your schema and generates strict anonymization configs automatically catching edge cases static rules miss. Safe staging with real data, minus the manual config work. Blog post: xata.io/blog/smarter-s…

Blog post here: xata.io/blog/database-…

The fastest way to make AI agents useful isn’t more tooling, it’s guardrails. Give @AmpCode a playbook: create a Xata branch, wait-ready, use the branch URL, run real queries, validate the fix. Video ↓ and the full write-up (AGENTS.md + PLAYBOOK.md pattern) in the thread.

Column constraints and table constraints look different in SQL but internally Postgres treats them the same. Both become rows in pg_constraint and both can be dropped using ALTER TABLE … DROP CONSTRAINT. @apatheticmagpie walks through clear examples and queries that show exactly how they’re represented. 👉 xata.io/blog/constrain…

Batching is a common pattern for processing large volumes of data but the right batch size depends heavily on the network setup. We added automatic batch size tuning to pgstream to adapt Postgres snapshots across different network environments. Here’s how it works and when it helps 👇 xata.io/blog/postgres-…

PostgreSQL stores all constraints: check, not-null, PK, FK, unique, exclusion and domain constraints as rows in pg_constraint catalog. In Postgres 18, even NOT NULL constraints now get their own entries here (before 18 they lived in pg_attribute!). If you want to understand how constraint enforcement works internally, @apatheticmagpie breaks it down beautifully in her latest blog: 👉 xata.io/blog/constrain…

Your database bill: $3,000/month But the database usage: - Prod: Active 24/7 ✓ - Staging: Used 2 hours/day ❌ - Dev branches: Used 4 hours/day ❌ You’re paying for 24/7 compute on databases that run maybe 10–20% of the time. That's $2,000/month on idle resources. Xata's scale-to-zero: → Active branch: Full compute → Inactive branch: Hibernated (you pay for storage only) → Access branch: Auto-wakes in seconds One customer cut their dev database costs by ~45% after turning this on. Stop paying for databases you're not using. Learn about scale to zero: xata.io/documentation/…

The feature: Automatic PII anonymization The payoff: Test with production-like data, without exposing PII Both security and speed win. Try it here: xata.io/get-access

One customer told us: "We demo to prospects with realistic financial data now. Fully anonymized. Changed our close rate."

The security vs. speed question: Can't test with production data → security risk Can't test with mock data → bugs in production Most teams choose one and lose on the other. Here’s how to get both security and realistic testing 🧵