Random performances on the “serverless” formula


Yesterday I switched from the standard formula with pods to the new “serverless” solution and I noticed that the performance at the moment is really random on queries (with a filter by metadata). I have a lot of latency and sometimes even timeouts, including in the Pinecone dashboard when I do a query: Query failed. Error: We were unable to process your request. If the problem persists, please contact us at https://support.pinecone.io

I am therefore hesitant to stay on serverless or to return to the traditional formula, the financial dimension is important for my project and the “serverless” solution therefore seems attractive, but I must also be certain of having reliable performance. I need to go to production mode in a few days, so it’s the last dev time to take the right decision.

Thank you for sharing your feedback :wink:

1 Like

Hi @Antoine, apologies for the delayed response. If you are still experiencing varying performance quality on your serverless indexes, I encourage you to share the structure of your metadata filters on this thread. Perhaps we can uncover what the root cause may be with that information.

That said, I must remind you that we do not recommend using serverless in production during the public preview period unless you have performed thorough testing and validated your use case, which it sounds like you have been doing.

Let us know if you have any questions!

Hi, it works much better now, maybe it was a little bug during the 1-2 first days you launched it. My app is still under dev but I think I can definitely rely on the serverless formula for production environement. Thanks for that great service it really change life to not pay for idle server time :wink:

1 Like

This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.