Occasional 429 while working #1987

earl-warren commented 2025年06月16日 14:42:14 +02:00

Comment

In the past weeks my work has been disrupted by 429 errors in ways that are difficult to figure out. I'll collect data to help figure that out here.

image

16 June - IP 87.88.168.40

• Noticed when posting a comment
  image
• https://codeberg.org/Codeberg/Community/issues/new => 429
• After a few minutes it was back to no 429

### Comment In the past weeks my work has been disrupted by 429 errors in ways that are difficult to figure out. I'll collect data to help figure that out here.  ## 16 June - IP 87.88.168.40 - Noticed when posting a comment  - https://codeberg.org/Codeberg/Community/issues/new => 429 - After a few minutes it was back to no 429

martin_schilliger commented 2025年06月26日 05:13:51 +02:00

Do you have an action runner? Mine always triggers the 429 error, it's rather annoying.
#1645

Do you have an action runner? Mine always triggers the 429 error, it's rather annoying. https://codeberg.org/Codeberg/Community/issues/1645

rfc2549 commented 2025年06月29日 04:41:05 +02:00

@martin_schilliger wrote in #1987 (comment):

Do you have an action runner? Mine always triggers the 429 error, it's rather annoying. #1645

I have been experimenting with an action runner and have been seeing 429's as well.
The runner and myself share the same public IP and in aggregate it is probably easier to hit the treshold for rate limiting.

The rate-limiting wears off eventually but in the meantime it means jobs are timing out/failing.

@earl-warren shared the 429 from the browser, from the runner's perspective it looks like this:
Screenshot From 2025年06月24日 22-28-29.png

I do not mean to be complaining, and I appreciate the difficulty of being scraped by crawlers, bots and LLMs.
Are there things we can do to alleviate the issue ?

It seems to trigger easily enough when the runner is running jobs and I am watching the jobs logs play out. Is this expensive ?
I'm having a hard time quantifying the amount of traffic from a client perspective. Should there be a slower polling ?

@martin_schilliger wrote in https://codeberg.org/Codeberg/Community/issues/1987#issuecomment-5605529: > Do you have an action runner? Mine always triggers the 429 error, it's rather annoying. #1645 I have been experimenting with an action runner and have been seeing 429's as well. The runner and myself share the same public IP and in aggregate it is probably easier to hit the treshold for rate limiting. The rate-limiting wears off eventually but in the meantime it means jobs are timing out/failing. @earl-warren shared the 429 from the browser, from the runner's perspective it looks like this:  I do not mean to be complaining, and I appreciate the difficulty of being scraped by crawlers, bots and LLMs. Are there things we can do to alleviate the issue ? It seems to trigger easily enough when the runner is running jobs and I am watching the jobs logs play out. Is this expensive ? I'm having a hard time quantifying the amount of traffic from a client perspective. Should there be a slower polling ?

rfc2549 commented 2025年07月22日 04:38:34 +02:00

I don't know if anything has changed in the past couple weeks but I have anecdotally not gotten as many 429's (or at all?) in my recent work sessions. I may have been more careful about less watching of the live job action log scrolling and the forgejo runner seems to be chugging along.

I don't know if anything has changed in the past couple weeks but I have anecdotally not gotten as many 429's (or at all?) in my recent work sessions. I may have been more careful about less watching of the live job action log scrolling and the forgejo runner seems to be chugging along.

earl-warren commented 2025年07月28日 21:21:03 +02:00

That no longer happens. Whatever it was is now resolved ❤️

That no longer happens. Whatever it was is now resolved ❤️