Python Durable Functions support in Azure Functions is now generally available
公開日: 3月 10, 2021
Durable Functions, an extension to Azure Functions that lets you define stateful serverless workflows, now supports Python in general availability. Developers and data scientists can now orchestrate complex data processing and data science workloads in Azure Functions using familiar Python language constructs.
A common scenario enabled by Durable Functions is chaining together a sequence of functions to build a long-running data processing pipeline. You write an orchestrator function that uses ordinary Python code to compose a workflow that calls other Azure Functions. Durable Functions takes care of checkpointing progress and retry logic. You can handle errors in your orchestration with standard try/except statements.
Another scenario made possible by Durable Functions is fanning out machine learning workloads to perform parallel tasks taking advantage of the elastic scale offered by Azure Functions’ serverless platform.
During the preview period, there was strong demand from customers for Durable Entities to be supported in Python. Starting today, Python developers can also build Durable Entities—stateful functions that provide an actor-like programming model.
See the documentation for more patterns supported by Durable Functions. To get started with Python Durable Functions, try the quickstart.