Add language that warns users about costs

[larsvilhuber]

https://github.com/grantmcdermott/codespaces-r2u/blob/dee1a42e3607b7709af8a4979c2cbc1e605198bd/README.md?plain=1#L34

Even closed, Codespaces still accumulate costs based on the storage used until they are deleted. While this is typically pennies, it's worth warning users about this (say, if you downloaded 500gb of data...)

[eddelbuettel]

Interesting. I wondered about that too. Then again my (test) ones got deleted quickly and I still seem to be in good standing with my github 'pro' account (and I never know whether that is due to $work having an enterprise account, or my part-time .edu involvement does that, or whether it is because I can't stop doing 'open sauce'). But net-net I only needed to add a credit card and accumulate some pennies in compute debt when I looked into the (truly expensive) arm64 runners for CI.

So yes, we should mention this.

[larsvilhuber]

https://docs.github.com/en/billing/managing-billing-for-github-codespaces/about-billing-for-github-codespaces

There's a free tier - 15GB-months storage, 120 core hours for free users. An empty container freshly started from this devcontainer has 21GB. So if you use this often enough in a month, you'll exhaust your free tier. The default minimal container has 2 cores, so at most 60 hours of usage (including any subsequent idle time).

Compute costs are until the container times out. Usually, that's 30 minutes https://docs.github.com/en/codespaces/customizing-your-codespace/setting-your-timeout-period-for-github-codespaces but can be configured.

And if you decide to pre-build (for faster start ups), you get charged for that too.

Most users won't incur any costs. But if you actually use it often, or with larger systems, you might.

[larsvilhuber]

Actually, only looks like an empty codespaces-r2u only incurs about 2GB - maybe the rest is mapped onto shared spaces? df -h . says it's using 21GB...

[eddelbuettel]

Ah yes, I had looked into this pricing info too. It is not too dissimilar from RStudio Cloud or Google Colab: some hours free, eventually you may. We used to have a three-year pilot with unlimited RStudio Cloud for students at U of Illinois, that is gone and we provide a (large enough) (single-server) RStudio instance. I am planning to list these as possible alternatives.

For real deployment, and at scale, it is a different story,

[grantmcdermott]

Thanks, agree.

@larsvilhuber do you mind putting in a PR with a sentence (or two) to that effect, with links, in the README? I'll be happy to merge.

[larsvilhuber]

Will do.