Persistence as it relates to Always On

I have an Always On repl that scapes waiver information from airlines’ alert websites and with a celery beat schedule, does so every 15 minutes. Before it scrapes the latest information, it copies output/latest.csv to output/previous.csv; it then writes new data to output/latest.csv. I then turn this into various data frames to understand what alerts have been added, dropped or changed. I’d appreciate any help in understanding what my risk/exposure is as it relates to this limitation:

Changes your program makes to the filesystem won’t persist, so we suggest you use Replit Database rather than, for example, writing to text files.

I understand Always On repl occasionally restart and gather the matter of persistence has to do with how it handles this (though haven’t discounted it might impact the program anytime I close the tab or let my computer go to sleep). My concern is that anytime the repl restarts that both output/latest.csv and output/previous.csv will be wiped [and when the program compares latest to previous, it will think everything is a new alert].

Really appreciate any assistance!

My understanding is a Repl’s filesystem is not persistent unless the Repl is hosted somehow. The hosted content is irrelevant and should prevent the loss of data.

Repl filesystem is persistent these days

1 Like

Ah, I see the Nov 2021 blog about file persistence that presumably supersedes the Feb 2021 blog I quoted. Much appreciated!