Surviving pgvector in Production: A Reality Check
Friday, June 26 at 11:00–11:50
-
As a Database Specialist at Google, Miguel helps developers and customers to build scalable and reliable applications on PostgreSQL. With a foundation of 10 years in Oracle Consulting, he has developed a deep understanding of the core challenges in database migration, performance, availability, and scalability. Now, his daily work involves everything from guiding enterprise migrations and fixing replication bottlenecks to sharing practical strategies on how PostgreSQL is evolving to support new frontiers, including the next generation of AI applications. In his personal life, Miguel enjoys outdoor activities, especially running, trail running, and hiking with his family
Building a vector search prototype is easy; keeping it running in production is a completely different problem. When developers push millions of high-dimensional embeddings into PostgreSQL, the physical limits of the database hit hard. Massive vector columns trigger heavy TOAST usage, evict critical data from shared_buffers, and cause storage footprints to explode, usually leaving DBAs to fix the resulting performance degradation.
Using data from real-world experiences, we will expose what actually happens to your infrastructure and query plans when vector tables grow exponentially and analyze the trade-offs between space utilization, query latency and recall.
Finally, we will explore how to shrink the storage and RAM footprint using halfvec and binary quantization, while measuring the cost to search accuracy.
