Given that we expect multi-billion vector counts, I think this (and potentially other areas throughout the code?) need to be updated to int64. https://github.com/pinecone-io/pinecone-java-client/blob/e5f79bb81e1d1748f4f6ff13c66d10442440c309/src/main/java/io/pinecone/proto/DescribeIndexStatsResponse.java#L192