Optional
Maximum number of concurrent API requests to the provider. Helps avoid rate-limit errors on shared-tier plans.
Hard ceiling on requests per second. The client will throttle to stay under this limit. 0 = unlimited.
Maximum number of vectors in a single upsert batch. Pinecone recommends ≤100; other providers may differ.
Maximum number of concurrent API requests to the provider. Helps avoid rate-limit errors on shared-tier plans.