Scaling Postgres

In this episode of Scaling Postgres, we discuss experiments to achieve four million transaction per second, the importance of extended statistics, parallelism in Postgres and an introduction to window functions. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/324-four-million-tps/ Want to learn more about Postgres performance? Join my...

In this episode of Scaling Postgres, we discuss a breaking change in the new version of PgBouncer, PostgreSQL 17 Beta 2 is released, examination of the new built-in collation provider in PG 17 and Notion's data lake. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/323-pgbouncer-breaking-change/ Want to learn more about Postgres perfor...

In this episode of Scaling Postgres, we discuss an incremental sort instability issue with the Postgres planner, whether we should use foreign keys, how the visibility map works and how to vacuum the template0 database. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/322-postgres-sort-instability/ Want to learn more about Postgres per...

In this episode of Scaling Postgres, we discuss the shutdown of Ottertune, how schema changes cause locks and how to avoid them, the benefits of on conflic do nothing, and pgvectorscale. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/321-ottertune-is-dead/ Want to learn more about Postgres performance? Join my FREE training called Po...

In this episode of Scaling Postgres, we discuss three organizations scaling their databases to 100 TB and beyond, collation speed, configuring memory and new AI extensions To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/320-100tb-and-beyond/ Want to learn more about Postgres performance? Join my FREE training called Postgres Performan...

In this episode of Scaling Postgres, we discuss a time when Postgres development stopped, two new extensions pg_lakehouse & pg_compare and the upcoming event Posette. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/319-when-postgres-development-stoppped/  

In this episode of Scaling Postgres, we discuss all the new features in Postgres 17 Beta 1, some features that did not make it, database collations & sorting and causes of slow commits. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/318-postgres-17-beta-1-released/  

In this episode of Scaling Postgres, we discuss a new time-series open source extension called pg_timeseries, Postgres ignoring indexes, JSONB selectivity issues, and geographically distributed multi-tenant applications. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/317-time-series-open-source-extension/  

In this episode of Scaling Postgres, we discuss new Postgres releases, optimizing a query to be 1,000 times faster, custom vs. generic plans and the pgtt extension. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/316-new-releases-1000-times-faster-query/  

In this episode of Scaling Postgres, we discuss how hacking on Postgres is hard, a notifier pattern for using Listen/Notify, using histograms for metrics and saturated arithmetic. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/315-hacking-on-postgres-is-hard/  

In this episode of Scaling Postgres, we discuss performance improvements for pgvector 0.7.0, a guide to vector embeddings, building a Retrieval Augmented Generation app and only allow logins to replicas. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/314-150-times-faster-pgvector/  

In this episode of Scaling Postgres, we discuss whether the future of MySQL is Postgres, how to use recursive CTEs, work on OrioleDB, and PG17 performance improvements. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/313-the-future-of-mysql-is-postgres/  

In this episode of Scaling Postgres, we cover potential features in Postgres 17 such as explain serialize, verbose copy, pg_buffer_cache_evict, as well as many others. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/312-postgres-17-commit-orama/  

In this episode of Scaling Postgres, we discuss a 1,400 times faster max and group by implementation, new quantization capabilities in pgvector, adaptive radix trees and splitting & merging partitions in PG17. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/311-max-group-by-performance/  

In this episode of Scaling Postgres, we discuss whether Postgres will pull a Redis, remembering Simon Riggs, built-in collation provider and C.UTF-8 in PG 17 and health checks. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/310-andres-microsoft-postgres-save-linux/  

In this episode of Scaling Postgres, we discuss whether Postgres will pull a Redis, remembering Simon Riggs, built-in collation provider and C.UTF-8 in PG 17 and health checks. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/309-will-postgres-pull-a-redis/  

In this episode of Scaling Postgres, we discuss high availability's future, a custom sharding solution by Figma, sharding pg_vector queries, and PG17 logical replication failover. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/308-scale-through-sharding/  

In this episode of Scaling Postgres, we discuss whether we should be using bytewise or linguistic ordering by default, how to transform data in Postgres, benefits of a transaction_timeout and how to enforce join order. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/307-collation-conundrum/  

In this episode of Scaling Postgres, we discuss one configuration change that resulted in an 11,000 times faster query, why Postgres is not using your index, backported PG improvements and parallelism with TOAST. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/306-scalability-limits-from-slru-lock-manager/  

In this episode of Scaling Postgres, we discuss one configuration change that resulted in an 11,000 times faster query, why Postgres is not using your index, backported PG improvements and parallelism with TOAST. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/305-11k-faster-configuration-change/  

In this episode of Scaling Postgres, we discuss new Postgres releases, performance comparisons between PgBouncer, PgCat and Supavisor, a new extension pg_analytics, and new planner capabilities in PG16. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/304-overhead-of-pg_stat_statements/  

In this episode of Scaling Postgres, we discuss new Postgres releases, performance comparisons between PgBouncer, PgCat and Supavisor, a new extension pg_analytics, and new planner capabilities in PG16. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/303-pgbouncer-pgcat-supavisor-fight/  

In this episode of Scaling Postgres, we discuss when and how you should use UUIDs or not, how to optimize space with column order and arrays, whether it is better to tune parameters or queries and what devs should know about Postgres with Andrew Atkinson. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/302-uuid-vs-bigint-battle/  

In this episode of Scaling Postgres, we discuss how you can build a GPT in 500 lines of SQL code, how to optimize extension loading, the best way to set passwords and being able to change generated columns. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/301-postgres-llm-os-30-times-faster-index-builds/  

In this episode of Scaling Postgres, we discuss how you can build a GPT in 500 lines of SQL code, how to optimize extension loading, the best way to set passwords and being able to change generated columns. To get the show notes as well as get notified of new episodes, visit:  https://www.scalingpostgres.com/episodes/300-postgres-celebration/