GitHub as Storage and Actions as Compute?
For a personal project I ended up using Github Actions to perform compute functions based on a cron schedule, and then commit the changes back to the repository: https://github.com/edelauna/stock-predictor/blob/prod/.github/workflows/cron-history.yml
As a solo contributor it seems to work ok, I was curious about potential drawbacks, and considerations I may want to consider if I wanted to start adding contributors, or merging dependabot PRs while a cron action was running heh. The dev ex kinda sucks since you're basically creating a janky CI system with artifacts, with the same inability to locally dev and test, but with way less documentation and ecosystem support (other users). For some simple script for internal use it's probably fine, but it's hard to build up from. The Actions ecosystem is so poorly documented and filled with a bunch of abandoned actions as libraries that don't work anymore. It's not fun to work in. I have a similar but minimal use case where some data is scraped and then committed into a repo via actions. I tried to modularize it so that's all that repo does. Another repo then fetches from that repo to actually console the data. If I needed to build something like this for actual production use, I'd probably use a serverless worker (Vercel or Cloudflare) and store it in a cloud KV instead. Link to example workflow: https://github.com/edelauna/stock-predictor/blob/prod/.githu...