Resource-based worker auto-tuning now in Public Preview.
New @workflow.init decorator allows __init__ to receive the workflow arguments
Announcing Worker Auto-Tuning. This feature allows Workers to scale up to the maximum bounds of CPU and Memory of the underlying compute node you are running Workers in.
The Start Delay workflow option is now available, allowing you to schedule future Workflow executions. Signals-With-Start bypass delays, while other signals are ignored. Set this option in the Workflow Options field of your SDK.
This release improves workflow management: safe eviction, customizable failure types. Adds HTTP CONNECT proxy support, worker client replacement, API key option. Float-based metrics, counting workflows optimized. LogRecord changes for better compatibility.
Python SDK v1.5.0
Schedules are in General Availability
Python SDK 1.4.0
v1.3.0 of the Python SDK includes dynamic Workflows, Activities, Signals, and Queries, the testing environment using the CLI, and Worker Versioning.