More controls for pgbouncer secrets configuration

[andrii-korotkov-verkada]

closes: #45171

Allow to disable adding default secret mounts for pgbouncer configs as well as metrics exported database url env variable. This can be useful for cases, where the value is retrieved other way, e.g. secrets provider class.

---

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in newsfragments.

[hussein-awala]

Could you add unit tests for this change?

[andrii-korotkov-verkada]

I've refined the functionality and added the tests.

[andrii-korotkov-verkada]

Also, if this gets merged, how can I release a new chart version?

[andrii-korotkov-verkada]

How do I trigger tests for Helm etc.? Are they automatically triggered after some time or just don't update the checks summary at first?

[andrii-korotkov-verkada]

I see, it has to be maintainer-approved. Curious if at least some tests can be made to run by default. That can greatly speed up the iteration.

[potiuk]

I see, it has to be maintainer-approved. Curious if at least some tests can be made to run by default. That can greatly speed up the iteration.

Not until your first PR gets merged. This is protection against bots mining bitcoins and roque PRs trying to exploit weaknesses in Github Actions, and it's mandated by the Infra and security team of The ASF.

[potiuk]

But you can run exactly the same tests locally -> All tests that are run in CI are runnable locally - that is basic assumptions for our CI system https://github.com/apache/airflow/tree/main/contributing-docs/testing -> you can see more details about all the types of tests we have and how to reproduce them locally 1:1

[andrii-korotkov-verkada]

Not until your first PR gets merged. This is protection against bots mining bitcoins and roque PRs trying to exploit weaknesses in Github Actions, and it's mandated by the Infra and security team of The ASF.

Ah, I see, that makes sense.

But you can run exactly the same tests locally -> All tests that are run in CI are runnable locally - that is basic assumptions for our CI system https://github.com/apache/airflow/tree/main/contributing-docs/testing -> you can see more details about all the types of tests we have and how to reproduce them locally 1:1

Thanks for the hints, I've naively tried to just run pytest with no args :)

[andrii-korotkov-verkada]

Tho I might be running into some incompatibility issues with pytest

ERROR: while parsing the following warning configuration:

  error::pytest.PytestReturnNotNoneWarning

This error occurred:

Traceback (most recent call last):
  File "/Users/andrii.korotkov/.pyenv/versions/3.10.12/lib/python3.10/site-packages/_pytest/config/__init__.py", line 1638, in parse_warning_filter
    category: Type[Warning] = _resolve_warning_category(category_)
  File "/Users/andrii.korotkov/.pyenv/versions/3.10.12/lib/python3.10/site-packages/_pytest/config/__init__.py", line 1677, in _resolve_warning_category
    cat = getattr(m, klass)
  File "/Users/andrii.korotkov/.pyenv/versions/3.10.12/lib/python3.10/site-packages/pytest/__init__.py", line 165, in __getattr__
    raise AttributeError(f"module {__name__} has no attribute {name}")
AttributeError: module pytest has no attribute PytestReturnNotNoneWarning

[potiuk]

Thanks for the hints, I've naively tried to just run pytest :)

It works as long as you have the same environment.

Those instructions on running tests are explaining how to get the SAME environment to run your pytest command in - in general. To avoid "works for me" (or rather "does not work for me").

Airflow has 700+ dependencies, 3 different databases in 8 versions, 10 integrations, kubernetes and a number of other variations - recreating it all by just running pytest is a nice dream (and our vision in the future with testcontainers and other things) - but so far adding tools to make things running in reproducible way is a bit more efficient (as long as you follow it).

[potiuk]

Tho I might be running into some incompatibility issues with pytest

Just follow the docs. Breeze is best to reproduce the env. You can also use uv for local venv. But the docs are complete and explain what to do.

[andrii-korotkov-verkada]

I've got non-db unit tests, helm tests and k8s tests to pass when running with breeze.

[potiuk]

I've got non-db unit tests, helm tests and k8s tests to pass when running with breeze.

Always Be Rebased.