throttling cpu on PI 4 causes jobs to constantly be missed

[tim-oe]

Things to check first

• I have checked that my issue does not already have a solution in the FAQ

• I have searched the existing issues and didn't find my bug already reported there

• I have checked that my bug is still present in the latest release

Version

3.10.4

What happened?

on a pi 4 2g
I have mariadb running locally with a single python process running
source project https://github.com/tim-oe/WeatherWatch

the following scheduling is working under default frequency:

self._scheduler.add_job(
            sensor,
            "interval",
            minutes=self._schedulerConfig.sensorInterval,
            name=SchedulerSvc.SENSOR_JOB,
            id=SchedulerSvc.SENSOR_JOB,
            coalesce=True,
            max_instances=1,
            replace_existing=True,
        )

        self._scheduler.add_job(
            pimetrics,
            "cron",
            minute=f"2-59/{self._schedulerConfig.piMetricsInterval}",
            name=SchedulerSvc.PI_METRICS_JOB,
            id=SchedulerSvc.PI_METRICS_JOB,
            coalesce=True,
            max_instances=1,
            replace_existing=True,
        )

        if Camera().enable is True:
            self._scheduler.add_job(
                camera,
                "cron",
                minute=f"3-59/{self._schedulerConfig.cameraInterval}",
                name=SchedulerSvc.CAMERA_JOB,
                id=SchedulerSvc.CAMERA_JOB,
                coalesce=True,
                max_instances=1,
                replace_existing=True,
            )

i then was trying to throttle down the idle frequency in
/boot/firmware/config.txt
arm_freq_min=300

then all jos started getting missed:
I added to both job_defaults and to each job misfire_grace_time=60*60

but still jobs are missed

How can we reproduce the bug?

on PI 4 throttle down the idle frequency in
/boot/firmware/config.txt
to
arm_freq_min=300

[agronholm]

Do I have to buy myself a PI 4 to debug this, or will you do it for me?

[agronholm]

What I mean to say is, how do yo know this is a bug in the scheduler, rather than the environment causing processes to stall? I have no way to reproduce this, so I don't think I can help you with what little clues you've provided here.

[tim-oe]

can you provide maybe logging names i can turn debug on that might give me more information i can provide that might shed some light? This is the only process that is running and other than when the scheduled processes run it sits mostly idle.
in increasing the min frequency to 500 the triggers fire.

i'm not sure if the grace period is set properly or how i can tell what the delay that's causing them to miss is. as to the amount of seconds to allow https://apscheduler.readthedocs.io/en/master/userguide.html#controlling-how-much-a-job-can-be-started-late

not sure where you're located but pi's aren't too costly, and with python being pretty much the de-facto language for tinkering with it, i bet there's a lot of projects leveraging your lib on it.

[agronholm]

can you provide maybe logging names i can turn debug on that might give me more information i can provide that might shed some light?

Here you go: https://apscheduler.readthedocs.io/en/3.x/userguide.html#troubleshooting