Metadata size limit 10kb or 40kb?

Support
1

In the documentation, I see that “Pinecone supports 40kb of metadata per vector.”

However while using the Python client, I received this error:
HTTP response body: {“code”:3,“message”:“metadata size is 14087 bytes, which exceeds the limit of 10240 bytes per vector”,“details”:}

I can work around this pretty easily, but I’m curious, am I making a mistake, or is there an inconsistency between the docs and the API?

2 Likes
2

Can confirm this as well. Encountered on a new index created today while using the Python client.

3

HI @alach11 and @hiranya911,

Can you confirm which environment you’re using for your indexes? Also, if you could post your index names, that would be helpful.

4

I encountered this on us-central1-gcp. I’m not able to share the index name on a public forum. But this was an instance set up via GCP marketplace.

5

I’m on us-central1-gcp and my index is called work-observations.

6

Thanks, both. I’ve escalated this to engineering.

7

Ok, you should both be good to go now. You’ll have to create a new index to take advantage of the increased metadata size; you can do that with your existing one so you don’t lost any data by making a collection first, then an index from that collection.

Sorry for the inconvenience.

1 Like
8

@Cory_Pinecone , I have the same problem with my huber-172d557.svc.us-east1-gcp.pinecone.io index:
{“code”:3,“message”:“metadata size is 10546 bytes, which exceeds the limit of 10240 bytes per vector”,“details”:}

Could you increase metadata size for me too?

9

Also have this error. Could possibly increasing metadata size also work for this account?

ApiException: (400)
Reason: Bad Request
HTTP response headers: HTTPHeaderDict({'content-type': 'application/json', 'date': 'Sun, 21 May 2023 23:06:10 GMT', 'x-envoy-upstream-service-time': '3', 'content-length': '115', 'server': 'envoy'})
HTTP response body: {"code":3,"message":"metadata size is 52074 bytes, which exceeds the limit of 40960 bytes per vector","details":[]}