Multigres Architecture | Scaling Postgres 418

In this episode of Scaling Postgres, we go over the architecture of Multigres (Vitess for Postgres), pgBackRest's status, pg_flight_recorder and general SQL optimizations.

Content Discussed

• Multigres Architecture Overview
• Two jobs, two processes: why Multigres has its own connection pooler
• Per-user pools that share fairly
• Pooling without choosing a mode
• One parse per query, no matter how many gateways
• pgBackRest Will Continue!
• Keeping pgBackRest Open, Healthy, and Community Driven
• pg_flight_recorder
• Why Giant IN Clauses Slow Down Your App
• Egress problems and where to find them
• PARTITION MERGE/SPLIT, Once More With Locking
• Welcome to ORDER BY jungle
• XID Wraparound's Equally-Evil Twin
• How we saved ~40GB of Postgres space wasted on indexes
• Data analyst vs width_bucket()
• Postgres in Production, Special Series: Deep Dive into pg_stat_statements (Part 3)
• I think AI can actually help me…
• CloudNativePG and Crunchy PGO: an honest, opinionated comparison
• Patch PgBouncer Today
• Eleven CVEs Walk Into a Release
• PostgreSQL 19 Beta: The Four Features You’ll Actually Feel
• Two Decades, Two RCEs: What pgcrypto Has Been Doing Since 2005
• Beyond Vector Search: Why PostgreSQL Could Become the Memory Layer for Enterprise AI Systems
• Table Access Methods Wake Up
• How to Cut Over After a PostgreSQL Migration
• PostgreSQL’s Growing Role in AI Infrastructure
• From Stream Processing to the "PostgreSQL Village": A Marketer's Journey to PGConf.de
• The True Cost of Database Optimization: Engineering Time
• From VMs to Kubernetes: A DBA's Journey in a Large Global Bank

YouTube Video

Podcast Audio