Features

• Run a Flask application with CherryPy via CherryFlask

• Schedule one-off, recurring, monthly, and on-demand jobs

• Run jobs in parallel with do_parallel()

• Expose a web-based task monitor with rerun and disable actions

• Persist scheduler state across restarts with filesystem or SQLAlchemy backends

• Support custom holidays calendars and timezones

Installation

pip install flask_production

Quick start

from flask import Flask
from flask_production import CherryFlask, TaskScheduler

def send_report():
    print("sending report")

app = Flask(__name__)
sched = TaskScheduler(check_interval=2)
sched.every("day").at("08:00").do(send_report)

cherry = CherryFlask(app, scheduler=sched)
cherry.run(host="0.0.0.0", port=8080, threads=5, debug=False)

When used with CherryFlask, the scheduler runs alongside the Flask app and is stopped cleanly when the server exits.

CherryFlask

CherryFlask wraps a Flask app with a CherryPy server.

CherryFlask(app, scheduler=None, silent=False, timeout=60)

Parameters:

• app (Flask): Flask application instance

• scheduler (TaskScheduler): optional scheduler to run alongside the app

• silent (bool): suppress request logging

• timeout (int): CherryPy socket timeout in seconds

from flask import Flask
from flask_production import CherryFlask

app = Flask(__name__)
cherry = CherryFlask(app)
cherry.run(host="0.0.0.0", port=8080, threads=5, debug=False)

TaskScheduler

TaskScheduler is the main class for defining and managing jobs.

TaskScheduler(
    check_interval=5,
    holidays_calendar=None,
    tzname=None,
    on_job_error=None,
    log_filepath=None,
    log_maxsize=5 * 1024 * 1024,
    log_backups=1,
    startup_grace_mins=0,
    persist_states=True,
    state_handler=None,
)

Parameters:

• check_interval (int): how often to check for pending jobs in seconds

  • default 5

• holidays_calendar (holidays.HolidayBase): calendar for schedules such as businessday

  • default US holidays

• tzname (str): timezone name used by default

  • default local timezone

• on_job_error (callable): callback invoked when a job fails

  • default None

• log_filepath (str): optional file path for rotating logs

  • default None

• log_maxsize (int): maximum size in bytes for the rotating log file

  • default 5 * 1024 * 1024

• log_backups (int): number of log backups to retain

  • default 1

• startup_grace_mins (int): grace period for jobs after a restart

  • default 0

• persist_states (bool): restore job logs and disabled state after restart

  • default True

• state_handler (BaseStateHandler): custom state backend

  • default FileSystemState()

Common scheduling patterns:

from flask_production import TaskScheduler

sched = TaskScheduler(check_interval=2)

# Run every minute
sched.every(60).do(my_job)

# Run once per day at a specific time
sched.every("day").at("08:00").do(my_job)

# Run every weekday at 08:00 in a specific timezone
sched.every("weekday").at("08:00").timezone("Europe/London").do(my_job)

# Run on the 31st of each month (strict_date=False allows the last day of shorter months)
sched.every("31st").strict_date(False).at("08:00").do(my_job)

# Run multiple times in a day
sched.every("day").at(["09:00", "17:00"]).do(my_job)

# Run a job in a separate thread
sched.every(30).do_parallel(my_job)

# Run a script from disk
sched.run_script("/path/to/scripts", "report.py", ["--daily"])

# Start the scheduler loop (blocking)
sched.start()

For standalone use, call sched.start(). When running with CherryFlask, pass the scheduler to CherryFlask(app, scheduler=sched) and let the app start it.

TaskMonitor

TaskMonitor adds a web interface for inspecting jobs, viewing logs, rerunning tasks, and disabling or enabling them.

TaskMonitor(
   app,
   sched,
   display_name=None,
   endpoint="@taskmonitor",
   homepage_refresh=30,
   taskpage_refresh=5,
   can_rerun=True,
   can_disable=True,
   enhanced_rerun=True
)

Parameters:

• app (Flask): Flask application instance

• sched (TaskScheduler): task scheduler with task definitions

• display_name (str): name of the application to be displayed

  • default app.name

• endpoint (str): URL endpoint where the monitor can be viewed

  • default "@taskmonitor"

• homepage_refresh (int): home page auto-refresh interval in seconds

  • default 30

• taskpage_refresh (int): task detail page auto-refresh interval in seconds

  • default 5

• can_rerun (bool): enable the rerun action on job pages

  • default True

• can_disable (bool): enable the disable/enable action on job pages

  • default True

• enhanced_rerun (bool): allow editing job arguments before rerunning

  • default True

from flask import Flask
from flask_production import CherryFlask, TaskScheduler
from flask_production.plugins import TaskMonitor

app = Flask(__name__)
sched = TaskScheduler(check_interval=2)

monitor = TaskMonitor(app, sched, display_name="My App")

sched.every("day").at("08:00").do(my_job)

cherry = CherryFlask(app, scheduler=sched)
cherry.run(host="0.0.0.0", port=8080)

The monitor is available at /@taskmonitor by default, or at the endpoint you pass to TaskMonitor.

State persistence

By default, scheduler state is persisted to a filesystem-backed state directory. You can also use a SQLAlchemy-backed store.

from flask_production import TaskScheduler
from flask_production.state import FileSystemState, SQLAlchemyState

sched = TaskScheduler(persist_states=True)
sched = TaskScheduler(persist_states=True, state_handler=FileSystemState())
sched = TaskScheduler(persist_states=True, state_handler=SQLAlchemyState("sqlite:///app_state.db"))

The SQLAlchemy backend requires sqlalchemy and sqlalchemy-utils to be installed.

Custom holidays and timezones

from flask_production import TaskScheduler
from flask_production.hols import TradingHolidays

holidays = TradingHolidays()
sched = TaskScheduler(holidays_calendar=holidays)
sched.every("businessday").at("10:00").do(my_job)

Examples and references

• The package includes tests covering scheduler behavior, plugin monitoring, and state persistence in the tests directory.

• The monitor exposes JSON endpoints for the full job list, a single job, and a summary view.

Example Gist here