To get the show notes as well as get notified of new episodes, visit:
https://www.scalingpostgres.com/episodes/390-1.2-million-messages-per-second/

In this episode of Scaling Postgres, we discuss how far Postgres can scale with queue and pub/sub workloads, temporal joins, IPC:SyncRep and nested partitioning.

Want to learn more about Postgres performance? 
Join my FREE training called Postgres Performance Demystified here: https://www.scalingpostgres.com/courses/postgres-performance-demystified/

Timestamps:

01:32 - Kafka is fast -- I'll use Postgres
08:37 - Temporal Joins
09:56 - Explaining IPC:SyncRep – Postgres Sync Replication is Not Actually Sync Replication
11:25 - Troubleshooting PostgreSQL Logical Replication, Working with LSNs
13:42 - Transaction pooling for Postgres with pgcat
14:29 - Returning Multiple Rows with Postgres Extensions
14:57 - Waiting for PostgreSQL 19 – Add psql PROMPT variable for search_path.
15:18 - PostgreSQL 19: Two nice little improvements: log_autoanalyze_min_duration and search_path in the psql prompt
15:55 - pgfeaturediff
16:18 - Consulting Corner
17:14 - Outro

#postgres #postgresql

Scaling Postgres

A Postgres-focused podcast episode covering a benchmark analysis showing Postgres achieving 1.2 million messages per second in pub/sub mode and competitive queuing performance, challenging the need for Kafka at typical workloads. Also covers temporal joins using window functions, nuances of synchronous replication (IPC syncrep wait events), troubleshooting logical replication conflicts with LSNs, PGCAT as a connection pooler alternative to PgBouncer, writing Postgres extensions that return multiple rows, upcoming PostgreSQL 19 features (search path in psql prompt, log_autoanalyze_min_duration), and a tool (pgfeatured.com) for comparing features across Postgres versions.

1.2 Million Messages Per Second | Scaling Postgres 390