Replies: 4 comments
|
💬 Your Product Feedback Has Been Submitted 🎉 Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users. Here's what you can expect moving forward ⏩
Where to look to see what's shipping 👀
What you can do in the meantime 💻
As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities. Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐ |
|
No, this use case is against GitHub Actions policies on two counts. |
|
I can't give you the official ruling — only staff can — but I'll be straight with you about where each half of this lands, because the two halves are not equally risky. Fetching market data hourly. The Additional Product Terms say GitHub-hosted runners are for building, testing, deploying and publishing software. A job that runs forever to collect data for your application is not that. Even with a 10-minute timeout and polite 429 handling, the purpose is data acquisition, and that's the part the language looks at. The Cloudflare challenge on your home server doesn't change it — it's a reason you want GitHub's IP reputation, not a software-development need. Using a branch as a data relay. This is the clearer problem. Repositories are for source and history, not for a rolling 30-day blob store. Rewriting a single commit on a dedicated branch every hour is exactly the pattern that makes repos heavy, breaks clones and mirrors, and pollutes caches. If anything in your design depends on "the repo holds the current snapshot," I'd change that regardless of what staff say. What I'd suggest instead: keep Actions for the build/test part only, and put the collection on infrastructure you own — a small VPS, a home server behind a tunnel, or a scheduled function on a cloud provider. Store bundles in object storage (R2/S3, or even the home server) rather than in a git ref. It also removes the artifact-retention question entirely. One thing worth noting about the answer you got above: "against policy on two counts" is a reasonable reading, but it's not a ruling, and only GitHub can tell you whether a narrow personal use like this is tolerated. If you do ask staff, ask about the two parts separately — the data collection and the branch-as-storage question have different answers in practice. |
|
I agree with the answers above, and I'd add a couple of points that might save you trouble. The riskiest part is the data relay on a branch. Rewriting a single commit with 36-110 MB every hour makes the repository grow, since old objects stay around until garbage collection, and using Git as temporary storage isn't what Actions or repositories are meant for. Since the real reason for hosted runners is the Cloudflare challenge on your home server, I'd look at fixing that part instead:
If you still want to use Actions for something, ask GitHub Support to confirm in writing, since only they can give an official ruling on the terms. |
Uh oh!
There was an error while loading. Please reload this page.
🏷️ Discussion Type
Bug
💬 Feature/Topic Area
Schedule & Cron Jobs
Discussion Details
I’m looking for policy guidance before enabling a scheduled workflow in a private repository. GitHub Support directed me to Community Discussions. I’ve read the Actions section of the GitHub Terms for Additional Products and Features and have two questions.
Would those terms allow a GitHub-hosted runner to fetch public market data from the SafeTrade API once an hour as part of my application’s normal operation? The collector would make up to 260 requests per run, including retries, with a 6-minute collection limit and a 10-minute job timeout. My home server receives a Cloudflare challenge, while GitHub-hosted runners received data in earlier tests. That difference is one reason I’m considering hosted runners. The collector stops on a challenge, HTTP 403, or HTTP 429 and honors Retry-After across runs.
May I use a dedicated branch of the same private repository to hold compressed data bundles until my home server downloads them? Each successful run would replace the branch’s single commit with a snapshot covering up to 30 days, estimated at 36–110 MB. If publication fails, the workflow would retain the unpublished data as an Actions artifact for 7 days. The long-term archive would be on my home server.
Could a GitHub staff member clarify how the policies apply to each part, or point me to published guidance that addresses this use case?
All reactions