UI crashes when filtering logs by log level

[rafagsiqueira]

Apache Airflow version

3.0.3

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

No response

What happened?

Getting "Minified React error #185; visit https://reactjs.org/docs/error-decoder.html?invariant=185 for the full message or use the non-minified dev environment for full errors and additional helpful warnings.", when filtering logs by log level.

What you think should happen instead?

Should see only logs filtered by that log level

How to reproduce

• Open a task instance
• Filter logs by log level

Operating System

Debian GNU/Linux 12 (bookworm)

Versions of Apache Airflow Providers

apache-airflow-providers-cncf-kubernetes==10.6.1
apache-airflow-providers-common-compat==1.7.2
apache-airflow-providers-common-io==1.6.1
apache-airflow-providers-common-sql==1.27.3
apache-airflow-providers-fab==2.3.0
apache-airflow-providers-microsoft-azure==12.5.0
apache-airflow-providers-odbc==4.10.1
apache-airflow-providers-smtp==2.1.1
apache-airflow-providers-standard==1.4.1

Deployment

Official Apache Airflow Helm Chart

Deployment details

Using kubernetes executor.

Anything else?

Happens also when clearing the log level filter.
Seems to be reproducible when logs are longer than 30.000 lines.

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

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template! If you are willing to raise PR to address this issue please do so, no need to wait for approval.

[eladkal]

Seems to be reproducible when logs are longer than 30.000 lines.

I am not able to reproduce this issue on main.

Tested with

from datetime import datetime
from airflow.decorators import dag, task


@dag(
    dag_id='log_generator',
    description='Generate 40K log lines',
    schedule=None,
    start_date=datetime(2024, 1, 1),
    catchup=False,
)
def log_generator_dag():
    @task
    def generate_40k_logs():
        """Generate 40K log lines"""
        for i in range(40000):
            print(f"Log line {i + 1}: This is a test log entry for performance testing")

    generate_40k_logs()


log_generator_dag()

[rafagsiqueira]

In my case it happens when there are many INFO log entries and one ERROR log entry. When I try to filter by the ERROR log level it crashes. Can you try that?

[eladkal]

In my case it happens when there are many INFO log entries and one ERROR log entry. When I try to filter by the ERROR log level it crashes. Can you try that?

Please provide minimal reproduce example

[rafagsiqueira]

This seems to do the trick. Always crashes when I try to filter by ERROR level.

from datetime import datetime
from airflow.decorators import dag, task
import logging
logger = logging.getLogger(__name__)

@dag(
  dag_id='log_generator',
  description='Generate 100K log lines',
  schedule=None,
  start_date=datetime(2024, 1, 1),
  catchup=False,
)
def log_generator_dag():
  @task
  def generate_100k_logs():
    """Generate 100K log lines"""
    for i in range(100000):
      logger.info(f"Log line {i + 1}: This is a test log entry for performance testing. Issues could be related to log wrapping of multiple long lines. Therefore I am adding a really long line here. Hopefully it will wrap and crash.")
    logger.error("Error log entry")

  generate_100k_logs()

log_generator_dag()

[jroachgolf84]

Are your Tasks actually outputting 100,000 lines of logs? Why is that the case? Can you reduce the logging?

[rafagsiqueira]

@jroachgolf84 I have tasks that output many lines of log using different log levels. The purpose of logging using different log levels is to be able to debug our dags. Reducing the logging seems to beat the purpose of using log levels in the first place. And 100,000 lines of logs is even reasonable, since we are talking about ETLs.