Cancelling a query from a UI client is more nuanced than it might seem. Here's how we implemented safe KILL QUERY operations in Tinybird.

We recently refactored the Tinybird frontend. Take a look behind the scenes at how we started, the choices we made, and how we arrived at a faster and simpler app deployment.

Building modern docs begins with a modern tech stack. Here’s how we rebuilt our documentation and what it enables us to do for our customers.

ClickHouse schema migrations can be challenging even on batch systems. But when you're streaming 100s of MB/s, it's a whole different beast. Here's how we make schema changes on a large ClickHouse table deployed across many clusters while streaming… without missing a bit.

Here's how we fixed a problem that had stumped us for over a year and saw a 5x performance improvement in the process.

A story about the Python GIL, threading, and fixing code that already works.

We recently introduced a pull request to ClickHouse that enables simple JOIN support for parallel replicas on ClickHouse. The solution may be simple and naive, but the ceiling for performance on distributed queries just got WAY higher.

Cloud infrastructure costs are going up for everyone, but expectations to perform haven't receded. Here are 7 strategies that we use at Tinybird to keep our costs low as we innovate new ways to help our customers achieve more with less.

We made 3 big changes that have shortened our frontend CI execution time despite a massive increase in code size and test coverage.

Our GitLab CI pipeline was slow and flaky. Here’s how we used telemetry data, real-time analytics, and smart k8s configuration to fix it.

How we used rendezvous hashing to horizontally scale Kafka consumers to support hundreds of concurrent topics with fewer connections, thus lowering our infrastructure costs.

Splitting CSV files is a must when dealing with large, potentially larger than RAM, files. But how fast can it be?

The **unmodified** support message we sent to one of our clients outlining potential performance gains through Kafka compression

In this post we explain the data architecture, infrastructure and how we scale our real-time analytics service with ClickHouse.

Part of providing an outstanding level of service is to react quickly and professionally when things go wrong. This is an account of a recent service disruption and how we handled it.