Introducing the new open-source LLM Performance Tracker

Introducing the LLM Performance Tracker, an open-source tool to monitor the performance of your AI apps in real-time, built with:

Fork the repo to deploy -> tbrd.co/llm-tracker

Read the blog post to see how we built it -> tbrd.co/llm-tracker-post

Explore the live demo to try it out -> tbrd.co/llm-tracker-demo

Query timeout responses will not move to Status Code 429

In our March 7, 2025 changelog, we announced that query timeout responses would transition from status code 408 (Request Timeout) to 429 (Too Many Requests).

After further internal discussions and feedback from our users, we have decided not to proceed with this change. Query timeouts will continue to return 408, as they did before.

Explorations in Forward

Explore your data with Explorations. Interact with the Tinybird UI assistant to create SQL or time series nodes in your Pipes.

Explorations

New email notifications for credit-based plans

Customers on credit-based billing plans will now receive automated email alerts as they approach their credit limit:

  • 80% consumed: First notification.
  • 95% consumed: Final warning before depletion.

Improvements and bug fixes

  • [Forward]: The tb dev command now directly modifies the state tracked by the Tinybird Local environment, unifying the development workflow. Changes made or data appended during tb dev persist in your Tinybird Local environment. tb dev
  • [Forward]: tb create no longer requires Tinybird Local to be running to scaffold a new project. tb create
  • [Forward]: Fixed an issue with Kafka Avro ingestion. We also rewrote the Kafka Connector documentation to make it clearer and easier to follow.
  • [Forward]: You can now deploy projects that only remove Connections (S3, GCS, or Kafka) from Data Sources. This previously caused deployment failures.
  • [Forward]: Deployment promotions now handle errors better (such as network issues), making your deployment process more reliable.
  • [Classic]: Fixed a bug where deselecting new columns while appending data to a Data Source did not work as expected.