"Connection not found" error in KubernetesPodOperator logs

[kosteev]

Apache Airflow version

3.1.0

If "Other Airflow 2 version" selected, which one?

No response

What happened?

Using KubernetesPodOperator in Airflow 3 generates some not-useful warning/errors.

While running KubernetesPodOperator and using default kubernetes connection id "kubernetes_default", the Kubernetes hook will fallback to cluster-derived credentials:

airflow/providers/cncf/kubernetes/src/airflow/providers/cncf/kubernetes/hooks/kubernetes.py

Line 183 in 767ae70

return Connection(conn_id=cls.default_conn_name)

This works as expected, but generates warning/error messages - see screenshot above.

What you think should happen instead?

No response

How to reproduce

1. Run Airflow 3 instance
2. Deploy DAG

import datetime

from airflow import models
from airflow.providers.cncf.kubernetes.operators.pod import KubernetesPodOperator

with models.DAG(
    dag_id="kpo_simple_echo",
    schedule=None,
    start_date=datetime.datetime(2000, 1, 1),
) as dag:
  KubernetesPodOperator(
      task_id="kpo_simple_echo",
      name="kpo-simple-echo",
      cmds=["bash"],
      arguments=["-c", "echo ZZZZZZZ"],
      image="gcr.io/gcp-runtimes/ubuntu_20_0_4",
      config_file="{path to kube config}",
  )

3. Observe logs

Operating System

linux

Versions of Apache Airflow Providers

No response

Deployment

Google Cloud Composer

Deployment details

No response

Anything else?

No response

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[HsiuChuanHsu]

I was able to reproduce the log with the DAG you provided.

I think this log message is still useful for diagnostics, but its current severity is too high. It shouldn't be logged as an error.

I'll submit a PR shortly to lower its logging level.