In this version, we created the pipeline deployments from scratch by introducing a new fault-tolerant and reliable architecture. Now enterprise users can run pipelines in production in a separate computation cluster that gives better performance and stability.
Also, system administrators can set up rate limits and retries to have better reliability and stability of deployed pipelines under the load.
We've updated the way how you start adding data to the canvas. We've created a new flow of connecting your data source and changed the look and feel of the data previewer.
Bricks "encoding", "missing values treatment", "parse list" and "parse JSON" became 5 times faster. Also, we've implemented more than 50 stability fixes for the other bricks.