Hi all! The year is winding down for many but Prefect is just getting started 😎 - here're just a few of the things we've been working on for the last few weeks.
The Run Timeline is a new way to explore your flow runs from the UI in real time. Get at-a-glance insight into your flow run state transitions, task run states and durations, and task parallelism, all in one place.
Building off the success of the Dashboard view, we expect the new Timeline to provide insights that were otherwise hard to come by - quickly spot infrastructure latency, Lazarus rescheduling events, and much more!
Helm Chart ⚠️ (experimental)
Our resident Cloud engineer Michael Adkins has been hard at work extending support for Kubernetes deployments of Prefect Server, including shipping an optional K8s Agent with the cluster. The Helm chart is still very much under development (⚠️ so use at your own risk! ⚠️) and we're welcoming contributions, suggestions, and advice! You can find the Helm chart here.
Coming Soon: Python 3.9 🐍 ⚠️ (experimental)
Prefect Core now has (experimental) support for Python 3.9, released just a few short weeks ago! While there's no guarantee Prefect extras/third party packages will work, we'll be keeping the pulse on each and updating our integrations with them as they also begin supporting Python 3.9.
Check out everything Python 3.9 has to offer in the official release.
Task Library Additions
The Public Changelog would never be complete without a shameless plug for the Task Library. As a place for first-class integrations between Prefect and your favorite tools (including Prefect itself!), the Task Library features a growing list of 35+ integrations with hundreds of ready-built tasks for you to drop into your flow with little to no configuration.
In addition to a number of useful extensions, enhancements, and updates to existing tasks in the library, we've got an exciting addition for Python notebook users everywhere.
Ever wanted to use Jupyter notebooks in your flows? Prefect now integrates directly with Jupyter to execute notebooks, allowing outputs in either `JSON` or `HTML` formats; leverage your existing notebooks to clean, harmonize, and visualize your data, directly from your flow.
You can read more about the Jupyter
ExecuteNotebook task here.
Interested in contributing? Josh Meek wrote a great blog post to help you get started with contributing to the Task Library!