Migrated from Pinecone to Weaviate for a 1M vector store - honest report

We just moved our marketing content similarity search from Pinecone to Weaviate. Our store is about 1 million vectors from processed blog posts and whitepapers.

The migration itself was straightforward using the Weaviate client. My main question is about long-term costs and indexing strategy. Pinecone's pod-based pricing was predictable, but I'm still figuring out Weaviate's resource usage on our own cloud.

For those who've done this, what should I watch out for in production? I'm especially cautious about query performance with hybrid searches and filter-heavy use cases. Does the open-source setup really scale better for a mid-sized marketing team's budget?