How do I schedule an interval job with APScheduler?

Question

I'm trying to schedule an interval job with APScheduler (v3.0.0).

I've tried:

from apscheduler.schedulers.blocking import BlockingScheduler
sched = BlockingScheduler()

def my_interval_job():
    print 'Hello World!'
sched.add_job(my_interval_job, 'interval', seconds=5)
sched.start()

and

from apscheduler.schedulers.blocking import BlockingScheduler
sched = BlockingScheduler()

@sched.scheduled_job('interval', id='my_job_id', seconds=5)
def my_interval_job():
    print 'Hello World!'
sched.start()

Either should work according to the docs, but the job never fires...

UPDATE:
It turns out there was something else, environment-related, preventing the task from running. This morning, the task is working fine without any modifications to the code from yesterday.

UPDATE 2:
After further testing, I've found that 'interval' jobs seem to be generally flaky... The above code now works in my dev environment, but not when I deploy to a staging env (I'm using a heroku app for staging). I have other apscheduler 'cron' jobs that work just fine in the staging/production envs.

When I turn on DEBUG logging for the "apscheduler.schedulers" logger, the log indicates that the interval job is added:

Added job "my_cron_job1" to job store "default"
Added job "my_cron_job2" to job store "default"
Added job "my_interval_job" to job store "default"
Scheduler started
Adding job tentatively -- it will be properly scheduled when the scheduler starts
Adding job tentatively -- it will be properly scheduled when the scheduler starts
Looking for jobs to run
Next wakeup is due at 2015-03-24 15:05:00-07:00 (in 254.210542 seconds)

How can the next wakeup be due 254 seconds from now when the interval job is set to 5 seconds??

Answer 1

You need to keep the thread alive. Here is a example of how I used it.

from subprocess import call

import time
import os
from pytz import utc

from apscheduler.schedulers.background import BackgroundScheduler


def job():
    print("In job")
    call(['python', 'scheduler/main.py'])


if __name__ == '__main__':
    scheduler = BackgroundScheduler()
    scheduler.configure(timezone=utc)
    scheduler.add_job(job, 'interval', seconds=10)
    scheduler.start()
    print('Press Ctrl+{0} to exit'.format('Break' if os.name == 'nt' else 'C'))

    try:
        # This is here to simulate application activity (which keeps the main thread alive).
        while True:
            time.sleep(5)
    except (KeyboardInterrupt, SystemExit):
        # Not strictly necessary if daemonic mode is enabled but should be done if possible
        scheduler.shutdown()

Answer 2

I haven't figured out what caused the original issue, but I got around it by swapping the order in which the jobs are scheduled, so that the 'interval' job is scheduled BEFORE the 'cron' jobs.

i.e. I switched from this:

def my_cron_job1():
    print "cron job 1"

def my_cron_job2():
    print "cron job 2"

def my_interval_job():
    print "interval job"

if __name__ == '__main__':
    from apscheduler.schedulers.blocking import BlockingScheduler
    sched = BlockingScheduler(timezone='MST')

    sched.add_job(my_cron_job1, 'cron', id='my_cron_job1', minute=10)
    sched.add_job(my_cron_job2, 'cron', id='my_cron_job2', minute=20)

    sched.add_job(my_interval_job, 'interval', id='my_job_id', seconds=5)

to this:

def my_cron_job1():
    print "cron job 1"

def my_cron_job2():
    print "cron job 2"

def my_interval_job():
    print "interval job"

if __name__ == '__main__':
    from apscheduler.schedulers.blocking import BlockingScheduler
    sched = BlockingScheduler(timezone='MST')

    sched.add_job(my_interval_job, 'interval', id='my_job_id', seconds=5)

    sched.add_job(my_cron_job1, 'cron', id='my_cron_job1', minute=10)
    sched.add_job(my_cron_job2, 'cron', id='my_cron_job2', minute=20)

and now both the cron jobs and the interval jobs run without a problem in both environments.

Answer 3

How can the next wakeup be due 254 seconds from now when the interval job is set to 5 seconds??

It's simple: you have many pending executions as your most of the jobs didn't completed in the interval-window of time.

You could use the following parameters in order to sort this out:

 **misfire_grace_time**:    Maximum time in seconds for the job execution to be allowed to delay before it is considered a misfire 
 **coalesce**:  Roll several pending executions of jobs into one

To read more, check the documentation here.