This is a provider package for
docker provider. All classes for this provider package
airflow.providers.docker python package.
You can install this package on top of an existing Airflow 2 installation (see
for the minimum Airflow version supported) via
pip install apache-airflow-providers-docker
This release of provider is only available for Airflow 2.2+ as explained in the Apache Airflow providers support policy https://github.com/apache/airflow/blob/main/README.md#support-for-providers
Remove 'xcom_push' from 'DockerOperator' (#23981)
docker new system test (#23167)
Add 'device_requests' parameter to 'DockerOperator' (#23554)
Fix new MyPy errors in main (#22884)
Fix mistakenly added install_requires for all providers (#22382)
Correct 'multiple_outputs' param descriptions mentioning lists/tuples (#22371)
Avoid trying to kill container when it did not succeed for Docker (#22145)
Add Trove classifiers in PyPI (Framework :: Apache Airflow :: Provider)
Fixes Docker xcom functionality (#21175)
Fix docker behaviour with byte lines returned (#21429)
Add support of placement in the DockerSwarmOperator (#18990)
Fixup string concatenations (#19099)
Remove the docker timeout workaround (#18872)
Move docker decorator example dag to docker provider (#18739)
Add a Docker Taskflow decorator (#15330)
This version of Docker Provider has a new feature - TaskFlow decorator that only works in Airflow 2.2. If you try to use the decorator in pre-Airflow 2.2 version you will get an error:
AttributeError: '_TaskDecorator' object has no attribute 'docker'
Add support for configs, secrets, networks and replicas for DockerSwarmOperator (#17474)
Optimise connection importing for Airflow 2.2.0
Adds option to disable mounting temporary folder in DockerOperator (#16932)
[FIX] Docker provider - retry docker in docker (#17061)
fix string encoding when using xcom / json (#13536)
xcom_allis set to
False, only the last line of the log (separated by
\n) will be included in the XCom value
DockerOperator in version 2.0.0 did not work for remote Docker Engine or Docker-In-Docker case.
That was an unintended side effect of #15843 that has been fixed in #16932. There is a fallback mode
which will make Docker Operator works with warning and you will be able to remove the warning by
using the new parameter to disable mounting the folder.
Auto-apply apply_default decorator (#15667)
Due to apply_default decorator removal, this version of the provider requires Airflow 2.1.0+.
If your Airflow version is < 2.1.0, and you want to install this provider version, first upgrade
Airflow to at least version 2.1.0. Otherwise your Airflow package version will be upgraded
automatically and you will have to manually run
airflow upgrade db to complete the migration.
Replace DockerOperator's 'volumes' arg for 'mounts' (#15843)
volumes parameter in
was replaced by the
mounts parameter, which uses the newer
mount syntax instead of
Entrypoint support in docker operator (#14642)
Add PythonVirtualenvDecorator to Taskflow API (#14761)
Support all terminus task states in Docker Swarm Operator (#14960)
Updated documentation and readme files.
Remove failed DockerOperator tasks with auto_remove=True (#13532) (#13993)
Fix error on DockerSwarmOperator with auto_remove True (#13532) (#13852)
Initial version of the provider.