|
Föderation EN Do 22.05.2025 22:49:25 @Codeberg I've been rate limited (which is fine) and I'd like to make sure it doesn't happen in the future. Does having a couple of runner pinging for 2 different repos the offending bit? Or my playing back and forth with actions in two different the problem? Or both? Because I'd really do not want to be a bad codeberger, but if that requires adjustment on my side, I need to what to know how to change my workflow! Thanks! Medien: 1 |
|
Föderation EN Fr 23.05.2025 00:54:36 @fraud The Actions-specific endpoints should not trigger rate-limiting. However, depending on your workflows, they can use up a lot of requests (e.g. cloning several Git repos based on how many Actions you import or jobs you run in parallel). Does rate-limiting happen for all routes, or only specific ones? E.g. can you still access https://codeberg.org (landing page / dashboard) when some repository-URLs are throttled, or does everything result in the same message? |
|
Föderation EN Fr 23.05.2025 01:03:28 @Codeberg on my desktop I was on a pull request screen, but from the same IP my runner was hitting two job again and again as I was debuging the action. After I triggered the rate limit, I didn't do more testing to not trigger some more agressive rate limiting after that. I did increase my runner's inteeval in the settings, fetch_interval to 20 and report_interval to 50. |
|
Föderation EN Fr 23.05.2025 12:14:34 @fraud The larger the intervals, the less traffic for us, but I think it's not necessary. report_interval is maybe a little too high, as this is the time that updates the runner status and logs. You probably want it the opposite way: High fetch interval (because it only checks for new jobs) and lower report interval (so you can have live status and logs) |
|
Föderation EN Fr 23.05.2025 13:10:37 @Codeberg thank you for the feedback. Thank you again. |
