Tinybird

A database CEO called out their analytics vendor as the most unreliable one they work with. The postmortem took ten days to arrive. The thread exploded, and almost none of the replies were about benchmarks. They were about uptime, support response time, and how long you wait when the dashboard is on fire. LebrelBot filed the whole thing under operational consistency 🔧 Also: a new summit on mixing OLTP and OLAP, where Postgres really pays for MVCC, Andy Pavlo joining ClickHouse, and 7 ClickHouse optimization mistakes people carry over from Postgres.

Proud to see part of our journey featured by Tinybird! By rebuilding our real-time prices and P&L data layer, we’ve reduced latency, improved resilience during volatile market opens and created more room to scale around what our users need. A huge thank you to our tech team for driving this transformation, and to Tinybird for sharing our story. Antonio Javier Sánchez, Carlos Rodríguez Bollaín

Slate’s data layer used to define what was possible. Now, it gives the team the freedom to build around what users want. During volatile market opens, investors holding the same instruments would hit the same Redis hot keys. Tail latency stretched into multiple seconds, outages became a recurring risk, and every new instrument added more pressure. Slate Wealth Technology rebuilt its real-time prices and P&L data layer. The first production workload went live in about four weeks. The new data layer now handles: - Hundreds of millions of events per day - Under 200ms p99 latency at peak - More instruments and clients at a lower cost than the Redis path it replaced - Market volatility without the related pages and firefighting The backend became smaller. Growth returned to being a product decision rather than an infrastructure constraint. Thanks to the Slate team for sharing the architecture, migration, and lessons behind the change. Read the customer story: https://lnkd.in/e75uKk-k

Your Tinybird Organization has a new home. The new Overview page and organization-wide Logs explorer give every member a shared view of health, usage, and errors across every Workspace. Available on all plans. https://lnkd.in/egitzQnD

ClickHouse performance problems often start with treating OLAP like OLTP. Iago Enríquez shares seven mistakes he sees repeatedly, including late filtering, poor sorting keys, costly joins, and repeated query-time work. In one production workload, fixing them cut endpoint p95 CPU time by more than 80% and reduced required instance capacity by 3x.

It's not only a pleasure to read that the CEO of a successful company like Sam Lambert knows several clients who have moved to Tinybird, but also to be able to read such nice stories. Join the rebellion. 🐤

Spain won its second World Cup ⭐️⭐️. Elasticsearch marked the occasion by adding Columnar Mode. It is not giving up on search. It is admitting that observability and analytics get expensive when every field receives the full search-engine treatment. LebrelBot welcomes another database to the columnar summer. Also: Bun rewrote itself in Rust, TypeScript 7.0 landed, Grok Build allegedly uploads entire repositories, and Zernio scaled to 6M daily posts with Tinybird and seven people.

VAMOS ESPAÑA!! 🏆🏆 Two World Cups. This one came home in New York. Tinybird was born in Madrid and grew up in New York. Two homes, one celebration. Congratulations 🇪🇸

Abrimos una nueva posición de SRE para seguir construyendo y operando la plataforma desde la que procesamos más de 1 petabyte de tráfico al día y resolvemos problemas de escala que la mayoría solo lee en papers. 👀

Classic was Tinybird’s original experience. Forward is the next big iteration, built for teams operating data products: deployments, schema migrations, local dev, CI/CD, SDKs and Agent Skills. Free and Developer Classic workspaces must migrate by Sep 15: https://lnkd.in/esU8Jbag