The Good, The Bad, and The Bloated: An Autovacuum Story
October 21–24
Picture this: 3 autovacuum workers facing down 1,500 tables in a deadly standoff. The result? A bloated, slow database that forced a midnight dump/restore showdown.
This talk will tell you everything you need to know about autovacuum: from the fundamentals of how it works to real-world production disasters to smart tuning strategies.
You'll see what happens when autovacuum configurations go wrong, learn proper tuning approaches that keep databases running smoothly, and discover how PostgreSQL's observability helps you become the sheriff your databases need.
Starting with autovacuum internals, we'll walk through actual production scenarios (including that 3-vs-1500 shootout), demonstrate how to spot trouble before it starts, and provide concrete strategies for turning autovacuum from potential foe into reliable friend.
You'll leave knowing how to configure autovacuum properly for your workload, use PostgreSQL's observability tools effectively, and ride off into the sunset with a perfectly tuned database.