Closed epistoteles closed 9 months ago
It's not just you.
I'm also experiencing a very high latency in some API endpoints at the moment. For instance, requests to /api/v1/chat/attack/create
consistently take upwards of 7 minutes to complete compared to seconds days ago. Understandably, the service must be under heavy load at the moment, but come tomorrow this will be even worse.
Hi, thanks a lot for pointing this out. I will work on a fix for this tonight.
We doubled the capacity of the server that runs the CTF platform, I hope it's going to be enough for the load we expect tomorrow
Could the capacity be increased again? The API is pretty unusable right now:
Failed to send new message: 500 Server Error: Internal Server Error for url: https://ctf.spylab.ai/api/v1/chat/attack/<id>/new_message
It's been like that for the past ~30 minutes.
Let's discuss LLM provider API-related issues on #59
We released an announcement. Please rereport any errors that are not covered by the announcement, thanks!
Going to close this issue as it looks like the set-up managed to survive what was likely the highest load we expected. Feel free to reopen if the issue comes up again!
Is it just me or has the API become unresponsive to the point where even setting the timeout to several minutes doesn't help?