这是indexloc提供的服务,不要输入任何密码
Skip to content

Manual triggers picked by scheduler with delay #58327

@devgonvarun

Description

@devgonvarun

Apache Airflow version

main (development)

If "Other Airflow 2/3 version" selected, which one?

3.1.1

What happened?

There is always a delay of around 2 minutes between manual dag triggers from UI and scheduler picking up the dag despite only a configured heartbeat of 2s. KubernetesExecutor. On dag trigger the dag goes into Queued state and changes to Running state once the worker pod is spawned. The task is still queued at this point until the worker pod gets into the running state. But it takes 2 minutes for the worker pod to spawn.

What you think should happen instead?

Scheduler should not have large delays to spawn worker pod.

How to reproduce

Airflow in kubernetes installation. with KubernetesExecutor. Trigger a dag manually from the UI.

Operating System

Debian GNU/Linux 12 (bookworm)

Versions of Apache Airflow Providers

No response

Deployment

Official Apache Airflow Helm Chart

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

Metadata

Metadata

Assignees

No one assigned

    Labels

    affected_version:3.0Issues Reported for 3.0area:Schedulerincluding HA (high availability) schedulerarea:corekind:bugThis is a clearly a bugneeds-triagelabel for new issues that we didn't triage yet

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions