[Bug]: Scheduler show-schedule command updates the tasks each time it is run

[tjeerddie]

Contact Details

No response

What happened?

The command show-schedule updates the scheduled tasks instead of only showing the next run time:

It only seems to happen in the CLI command and the Graphql fetch that uses the same logic doesn't update the jobs.

discord question: https://discord.com/channels/1295834294270558280/1426204559985868895

Version

4.5.0

What python version are you seeing the problem on?

No response

Relevant log output

[Mark90]

Can confirm this happens on our implementation as well.

Additional hints:

• This only happens for apscheduler jobs using the "interval" trigger
• The apscheduler_jobs DB table contains a column next_run_time which has the correct time and is not altered when running show-schedule

Running the show-schedule command 2 times shows that the next run time changes each time:

[resume-workflows] Next run: 2025-10-13 16:12:32.276906+00:00 | Trigger: interval[1:00:00]
[clean-tasks] Next run: 2025-10-13 21:12:32.277081+00:00 | Trigger: interval[6:00:00]

...

[resume-workflows] Next run: 2025-10-13 16:14:13.499671+00:00 | Trigger: interval[1:00:00]
[clean-tasks] Next run: 2025-10-13 21:14:13.499851+00:00 | Trigger: interval[6:00:00]

This DB query shows that the actual run times are not updated are updated as well..

SELECT
    jobs.id,
    to_char(
        to_timestamp(jobs.next_run_time),
        'YYYY-MM-DD"T"HH24:MI:SS.MS"Z"'
    ) AS next_run_date
FROM apscheduler_jobs jobs
where jobs.id in ('clean-tasks', 'resume-workflows')

id,next_run_date
resume-workflows,2025-10-13T16:14:13.499Z
clean-tasks,2025-10-13T21:14:13.499Z

The CLI command should use the same logic as the GraphQL scheduledTasks query, without modifying the DB.

[hanstrompert]

Suggestion is to not start the scheduler, but get the schedules directly from the job store, if at all possible.
Time boxed to 3 points. Otherwise discus again.