Hit by rate-limiting

Hello,

I’m running a claude remote-control session in a VM in Scaleway.

I’m sending authenticated requests to drive the agent through codefloe issues, but since yesterday, I receive 429 Too-Many Requests errors.

It doesnt look like it was hitting very intensively the API when it started. I thought the rate-limit would reset, but 10 hours later, I’m still stuck.

Running the same request, with the same authentication token from my own laptop does work. I’m wondering if someone isn’t messing with another Scaleway VM and if I might be hitting a rate limit that isn’t coming from me.

I can send further details in a private channel.

Thanks for the report, this was a bug on our side and you don’t need to send anything privately.

It had nothing to do with other Scaleway VMs.
Our proxy treats issue and PR list calls with filters or sorting (labels=, sort= and so on) as anonymous unless they carry a browser cookie, and those are limited to 10 requests per 10 minutes per IP.

An API token didn’t count as logged in, so your agent’s polling hit that limit, and each rejected retry kept the counter above it.

Your laptop has its own IP and its own counter, which is why the same token worked there.
We have deployed a fix that exempts authenticated API requests from this limit. Let us know if this solved your problem!

Hello.

Thank you for your quick reply.

I confirm, I’m unblocked.