Skip to content

Releases: Azure/azure-functions-python-worker

Release 1.2.5

23 Jul 21:16
Compare
Choose a tag to compare
  • Fix an issue where the azure namespace is not initialized in placeholder mode
  • Disable dependency isolation to fix an issue where azure. namespace related libraries are not able to load in Linux Consumption

Release 1.2.3

04 Jun 21:52
Compare
Choose a tag to compare
  • Re-enable PYTHON_ISOLATE_WORKER_DEPENDENCIES.
  • Fix an azure.functions module issue in its _thirdparty folder and fix azure namespace discovery.

Release 1.2.2

02 Jun 20:22
Compare
Choose a tag to compare
  • Fix an issue where Python 3.9 Linux Consumption apps are facing 404 and 503 when the customer's files are not mounted during specialization.
  • Re-enable worker status response to optimize the throughput

Release 1.2.1

04 May 21:52
Compare
Choose a tag to compare
  • [HotFix] Disable WorkerStatus reporting due to the gRPC channel fails to acquire WorkerStatusResponse message from Python Queue structure, causing the Function Host /admin/host/ping endpoint to return 408 (timeout).

Release 1.2.0

05 Apr 19:25
Compare
Choose a tag to compare
  • Added a new mechanism to transfer data between Functions Host and Python Worker via shared memory. This feature will improve the data transfer efficiency and performance especially when moving byte arrays between Host and Worker (e.g. Blob bindings). This feature can be turned on by setting FUNCTIONS_WORKER_SHARED_MEMORY_DATA_TRANSFER_ENABLED to true (PR: #816, thanks @gohar94)
  • Added a new worker interface to allow third-party library integration into Python Worker. Third-party library developers can now integrate with Azure Functions by implementing lifecycle hooks in azure.functions.FuncExtensionBase or azure.functions.AppExtensionBase. This feature can be turned on by setting PYTHON_ENABLE_WORKER_EXTENSIONS to true. It is enabled in Python 3.9 by default. (PR: #815)

Release 1.1.12

18 Feb 23:40
Compare
Choose a tag to compare
  • Implemented capability to isolate worker dependencies (#794)
    • New Experiment feature flag introduced - PYTHON_ISOLATE_WORKER_DEPENDENCIES (default: False)
  • Setting PTPTC to None for Py3.9 only (#791)
    • More info on PYTHON_THREADPOOL_THREAD_COUNT in the docs page
  • Update Python SDK Version to 1.6.0 (#812)
  • Adding updates to the log messages (#805)

Release 1.1.10

24 Nov 09:31
Compare
Choose a tag to compare
  1. Add support for Python 3.9 on Windows

Release 1.1.9

18 Nov 21:34
Compare
Choose a tag to compare
  1. Reduce Linux and macOS Python worker size by enabling manylinux2010 wheel in Python worker. (#782)

Release 1.1.8

10 Nov 21:10
fd0a96b
Compare
Choose a tag to compare
  1. Add support for Python 3.9 on Linux and macOS. Windows is currently under development as gRPC Wheel for Python 3.9 is missing. [#751, #766, #775]
  2. Fix an issue where PYTHON_THREADPOOL_THREAD_COUNT is not recognized in Linux Consumption SKU [#774]
  3. Reverting the support for debug logs in executed functions introduced in #745. [#780]
  4. Add support for ServiceBus trigger to handle batch messages when "cardinality" is set to "many" in function.json [Azure/azure-functions-python-library#73]

Release 1.1.6

01 Oct 02:27
72e7c3e
Compare
Choose a tag to compare
  1. Update azure-functions library to 1.4.0
  2. Add PYTHON_THREADPOOL_THREAD_COUNT app setting (#744)
  3. Adding support for debug logs in executed functions. (#745)