Helicone Community Page

Updated 3 months ago

Probable issue with stats shown in cache

I've recently setup caching on helicone. In no time the cache hit stats have spiked. We don't even serve this much traffic. πŸ€”
I feel the stats are incorrect. Can someone please look into this?
Attachment
Screenshot_2023-09-20_at_11.00.11_PM.png
J
R
20 comments
looking now! thanks for flagging - is it okay if i add myself to your org?
Yup, that's fine./
Just to confirm these stats show cache hits and not number of requests that are cached, right? In either case, it seems to be completely off the charts if I consider the tradffic we are serving AI features rn.
We are getting actual cache hits from your service.

This is the request that is getting hit:
https://www.helicone.ai/requests?requestId=985603ba-a42f-42de-a5a1-047673400280

---

The graph shows cache hits.

You can see the all the cache hits that are coming in to our serve below. You might have a job that is running right now that keeps requesting the same thing over and over, would you mind checking your services?
Attachment
image.png
Thanks for looking into this. I'm checking it on my end.
(unrelated to this thread) Meanwhile can you please answer these:
  1. How/where to reach out to get information about storage consumption?
  2. How/where to reach out if I need to purge cache?
Do you have information about the user who is requesting these repeated requests?
I can not find it on the dashboard πŸ€”
Also, the total number of requests in the users dashboard (sum of all user's requests) way too less than the number of cache hits πŸ€”
I did check out services but couldn't find anything wrong there.
I feel there could be either of these issues:
  1. can be an attack
  2. some glitch on your end
  3. something went really wrong with our scripts
To identify the root cause, can I request more detail on the requests (request-id, user-id/email or any other info through which we can identify user) that ended up in repeated cache hits.
Hi
  1. About storage consumption, right now we will notify you if you are getting close. I will check your org for you, but most of our users are well below the 2Gb storage requirements
  2. Right now purging cache is not a feature.
We don’t not store that information.

We checked our services and we are getting requests from your service. I believe there is an issue with your scripts. The requests are not behaving like an attack.
Understood.

  1. Yes, please LMK the current consumption.
  2. We can not even request cache purge? Not as a feature on dashboard, but something that helicone support can do for us internally?
It seems like this has stopped
Yes, but we still don't know the root cause. Anyways, we can close this issue for now.
Although, there are 2 other things I requested above. Can you please response on them?
Yeah for Caching purge, we can support this internally for you. But we just added this issue to our active roadmap for you
https://github.com/Helicone/helicone/issues/859
Thanks a lot
What about the current storage consumption? You said I can request and you will LMK, so just want to check what is the current storage consumption out the 2 GB cap?
Hi Sure thing - I just put a request in. Take a little while to calculate
Sure, LMK by when can I expect this? I know you guys might be very busy with all the work. So, don't want to keep on pop up a message for the same thing.
Can you check your DMs?
Add a reply
Sign up and join the conversation on Discord