Seattle Postgres

PostgreSQL has become the go-to database for modern applications, but scaling beyond a single instance remains a hard problem. Multigres is an open-source project from Supabase that brings horizontal scalability and high availability to PostgreSQL by adapting the proven architecture of Vitess, the sharding system that came out of YouTube and now powers some of the world's largest MySQL deployments.

This talk covers how Multigres works: a two-level proxy architecture that presents a distributed cluster as a single Postgres server, intelligent query routing and connection pooling, and an orchestrator that handles failover automatically using unmodified PostgreSQL. Whether you're hitting the limits of a single instance today or planning ahead, you'll leave with a clear picture of where PostgreSQL scaling is headed.

Pizza is provided, and we also have a user group lending library of books & papers related to Postgres and Databases.

===== Speaker:
Peter Soderberg is a Customer Solutions Architect at Supabase helping users get the most out of Postgres and the ecosystem. He brings over a decade of field engineering experience, from large-scale analytics in Hadoop-land to real-time query engines on TB-per-hour+ Kafka streams. Prior to Supabase, he worked at Imply (Apache Druid) and Elastic helping organizations wrangle high-throughput pipelines, lightning-fast queries, and complex migrations. Peter lives in NW Seattle and has been a regular face at SEAPUG.

Join us for an evening with talks on two Postgres topics: AI Agent Skills and Change Event Streaming!

Pizza is provided, and we also have a user group lending library of books & papers related to Postgres and Databases.

===== Talks:
Savannah Longoria: Change How You Vibecode with Postgres

When you vibecode with your LLM of choice, your database layer is usually the weakest output: naive queries, missing indexes, no RLS, app-layer workarounds for things Postgres handles natively. The postgres-dba skill changes that. AI agent skills are specialized, portable packages of instructions, scripts, and resources - often defined by a SKILL.md file - that extend an AI's capabilities to perform specific, repeatable tasks. In this talk I'll show the before and after: same prompts, dramatically different output. Proper indexing strategies, window functions instead of app logic, safe migrations, replication setup, and performance diagnostics. I'll also share what I learned building and testing this skill: what works, what doesn't, and how to get the most out of AI-assisted database development.

.

Narissa Tsuboi: From WAL to Kafka - Turning Postgres Changes into Event Streams

If Postgres is where your data lives, Kafka is how that data travels. Change Data Capture (CDC) bridges the two by turning database changes into a stream of events that other systems can react to.

This talk explores how PostgreSQL logical replication enables CDC through WAL decoding and replication slots, and how the Debezium Postgres connector captures those changes and publishes them into Kafka topics. We will go through the key pieces of configuration on both the Postgres and connector sides, along with how this setup allows downstream systems to consume row-level changes without issuing queries against the database.

We will also cover common use cases, monitoring considerations such as replication lag and WAL growth, and performance tradeoffs observed in production. The goal is to provide a clear, practical understanding of how to move changes from Postgres into Kafka using CDC.

.

===== Speakers:
Savannah Longoria is a Senior Developer Advocate at Databricks. She's spent her career in the database world, from SingleStore to PlanetScale to Neon (now part of Databricks). She started as a Solutions Architect helping customers design schemas and optimize queries, and now focuses on helping developers build simpler, more powerful systems. Based in Seattle, WA, Savannah is passionate about making complex database concepts accessible and meeting developers where they are.

Narissa Tsuboi is a software engineer at Read AI, an AI notetaker and productivity platform, where she works on all things backend engineering, event-streaming, and data ingestion. She earned her Masters degree in Computer Science from Seattle University and prior to becoming a software engineer, she worked as a civil engineer for 5 years. Narissa loves to learn the internals of systems, how things go together, how they break, and then share that knowledge with others!