This repo is a template for Flask Heroku deployment of scheduled jobs. We often run into the need to run "cronjob" like jobs on a server fr our machine learning pipelines. There are many ways to do this, but deploying to Heroku is the easiest. This code is supposed to be a template for how to do that quickly.
There are two files of interest for this app, the myapp.py, which contains a dummy web app that just returns simple text. This must be there because Heroku needs a web process to run.
The other file, manage.py, is more interesting. This is where you can define commands to be run either on-demand or by the scheduler. This file is written using flask-script, a lightweight management plugin for Flask. We have defined only one command in there, hello. You can run it locally by typing:
python manage.py hello
If you have the Heroku CLI installed, you can type this instead:
heroku local:run python manage.py hello
This same job can be run on-demand on a deployed Heroku app by typing:
heroku run python manage.py hello
The scheduler uses this same interface to run. To use it, go to the web portal for your Heroku app and configure the Heroku Scheduler plugin for your app. Then just setup a job like below.
After running, you should see these lines in your Heroku logs. You can see these logs by simply typing:
heroku logs
