Consider switch from celery to BackgroundScheduler()

[aaron-schroeder]

I don't think I want celery taking up valuable ram just to support the one-time onloading of data. I would be perfectly fine waiting for the activities to be gradually added, as the server memory and Strava rate limits allow.

I think the cron job approach will be best, periodically checking for more activities and saving a few at a time, writing to a logfile somewhere. I can adjust the time between jobs or the number submitted based on rate limits. Or I could just handle the stravalib error.

The cron job can be done manually or using BackgroundScheduler().

References

• https://github.com/ethanopp/fitly/blob/8ee1c9db840b61424c0eea81b7805aa65f097146/src/fitly/app.py#L36-L60
• https://blog.miguelgrinberg.com/post/run-your-flask-regularly-scheduled-jobs-with-cron

[aaron-schroeder]

I ended up getting celery to work beautifully, and it will be the default approach to background tasks going forward.