Build
Containerisation of your workloads is a requirement for the Celery Executor installation type.
Note
See Makester's docker
subsystem for more detailed container image operations.
The container image is based on a customised variant of the Apache Airflow project's Dockerfile. The main difference is that the PYTHON_BASE_IMAGE
is overridden with Ubuntu as the underlying OS. This provides better currency to mitigate CVEs. The final Flowz Airflow image extends the customised base container image with the project's workflow capability (DAGs, operators and hooks).
Container image validation in local environment
Local container image build and validation allows you to launch a running instance of Apache Airflow before pushing to the image repository.
Note
Flowz embeds your workflow capability into the Flowz container image. As such, modifications to the workflow logic will require a fresh container image build.
Build the container image for local testing
The conatiner image build process will created two tags:
- the SemVer taken from
src/flowz/VERSION
latest
Note
See Makester's versioning subsystem on how Flowz maintains release versions.
Search for built container image
Typical output based on SemVer value of 2.8.1-0.1.1
contained within src/flowz/VERSION
:
REPOSITORY TAG IMAGE ID CREATED SIZE
loum/flowz 2.8.1-0.1.1 bcb2c5443e69 9 minutes ago 1.65GB
loum/flowz latest bcb2c5443e69 9 minutes ago 1.65GB