The rise of AI builders is causing an explosion in the number of provisioned Postgres clusters, creating an unprecedented management challenge that demands extreme automation. To navigate this new reality, we can frame the path to a self-driving Postgres using the six-level (0-5) autonomy model developed by the Society of Automotive Engineers (SAE) for self-driving cars. This talk will provide a high-level overview of 20+ key areas -- from cost optimization to data lifecycle management -- where automation is required to reach Level 4 and eventually Level 5.
While the end goal is full autonomy, achieving it starts with a solid foundation. We will argue that the most critical first step is building automation-friendly monitoring and automated health checks, which provide the ground truth for any autonomous action. From this foundation, we will deep-dive into two narrow, specific and "boring" challenges: first, achieving zero-downtime, fully reversible, zero-data-loss major upgrades; and second, the practical problem of proactive index health maintenance, demonstrated through the architecture of new tool pg_index_pilot. This session is for engineers building or managing large-scale PostgreSQL environments who are focused on solving the next set of practical operational hurdles.
Pizza is provided, and we also have a user group lending library of books & papers related to PostgreSQL and Databases.
