Back to Explore

Azure/azure-functions-durable-python

GitHub
1 updates · last 90 days1 watchersOpen source

Last release:

Python library for Azure Functions that provides Durable Functions bindings, enabling stateful serverless workflows. It helps you implement orchestrator and entity functions using the Azure Functions programming model, while handling state, checkpoints, and restarts. Useful for building resilient, stateful workflows in Azure Functions with Python.

Project status

  • Actively maintained: the most recent upstream push and update are on 2026-07-09 (v1.6.0), indicating ongoing work rather than a dormant state.
  • Apparent update cadence: updates arrived around every 4 to 5 months in the last few versions (v1.4.0 on 2025-09-24, v1.5.0 on 2026-02-04, v1.6.0 on 2026-07-09), suggesting a steady, evolving development rhythm.

AI summary generated

AI-generated from public sources. May be inaccurate. Report

Recent updates

  • v1.6.0

    Durable Functions for Python v1.6.0 introduces history-based replay detection, adds an orchestration restart API to DurableOrchestrationClient, and refactors payload serialization to go through a centralized df_dumps/df_loads shim. It also adds type-aware serialization/deserialization plumbing by threading discovered return types (and optional caller-supplied types) into the decoding path when the installed azure-functions SDK exposes centralized serializers.

    BreakingFeatures

  • v1.5.0

    Durable Functions for Python v1.5.0 adds support for explicitly overriding orchestration versions when starting new orchestration instances, and improves HTTP client behavior for better logging correlation and connection reuse. It also includes a documentation rename for the OpenAI integration and bumps aiohttp to a newer patch release.

    Features

  • v1.4.0

    This release adds an OpenAI Agents integration that is designed to run agent orchestration reliably on Azure Durable Functions. It introduces new Python modules under azure.durable_functions.openai_agents and adds a new DurableApp decorator that wires model invocations into durable activities with replay-safe task tracking.

    SecurityFeatures

  • v1.3.3

    This release adds a new `version` property to `DurableOrchestrationContext`, populated from the orchestration execution started event in the instance history. It also bumps the `aiohttp` dependency from 3.12.9 to 3.12.14. The code diff additionally introduces a new sample app demonstrating orchestration versioning.

    Features

  • v1.3.2

    Durable Functions for Python v1.3.2 includes a fix related to create_timer behavior when using non Azure Storage backends such as DTS and MSSQL. The release notes describe a long timer scheduling bug fix for those backends.

  • v1.3.1

    v1.3.1 updates the package to require Python 3.9 instead of Python 3.6 and refreshes dependency versions in setup.py. The repo CI workflow and contribution docs were also updated to reflect the Python version support change.

    Breaking

  • v1.3.0

    Release v1.3.0 adds long timer support, OpenTelemetry-based distributed tracing, and improved unit testing ergonomics for orchestrators, entities, and durable client functions. It also includes bug fixes around exception handling (including exceptions with no message) and task_any duplicate task results, plus dependency bumps and a Python version floor update.

    Features

  • v1.2.10

    Durable Functions for Python v1.2.10 includes fixes around orchestrator output serialization, entity id comparison, and HTTP content handling from orchestrations. It also adds repository-level development/CI configuration (devcontainer, validation workflow, CodeQL, and templates). The only likely developer-facing behavioral change beyond the release notes is around how call_http treats string content payloads by default.

    Features

  • v1.2.9

    Durable Functions for Python v1.2.9 adds client-side support for suspending and resuming orchestration instances via new REST calls. It also includes a fix so activity results of None are handled correctly during history parsing. In addition to the documented changes, it extends several public enums and adds new binding URL fields related to suspension and resumption.

    Features

  • v1.2.8

    This patch release updates the durable functions Python library to be compatible with upcoming `SettingsApi` changes in `azure-functions`. The code changes focus on adding conditional support for `SettingsApi` and adjusting the `Blueprint` class accordingly, along with a small test update.