HOW TO TAME A MASTODON: LESSONS FOR POSTGRESQL AT SCALE

Managing and running PostgreSQL can pose interesting challenges when you’re operating databases at a large scale. This talk will cover some of the things we’ve encountered in the last year working with a 25TB+ PostgreSQL database. Using an unnamed real life client, this case-study is a guide for problems encountered and solutions.

Scale environment we’ll use as a sample:

• 25TB database with significant growth

• 15 replicas

• WAL volume of 150GB/hr 

• Transaction volume of 230M+/hr on primary

We will cover 

• Big Table problems: how to add columns, constraints, and indexes with a huge single table in production

• Vacuum management for large databases

• High Transaction problems and solutions

• Scaling challenges with read replicas