More Data Centers, Fewer Problems

As Datadog continues to grow, their team needed to prioritize data center expansion. Unfortunately, their Postgres architecture—previously supporting a handful of data centers—became a painful liability for operators and service owners. Hidden coupling, operational toil, and reliance on components like PgBouncer surfaced major coordination challenges for data center expansion.

To understand what needed to change, the team used AI-assisted analysis to examine how Postgres was actually being used across hundreds of services. By analyzing real production workloads, queries, and traffic patterns, they identified hidden dependencies and unsafe assumptions that were impossible for individual teams to investigate alone, allowing them to deliver architectural and service-level changes with confidence.

In this talk, the speaker will share how their team simplified a production Postgres architecture to enable safe, repeatable, and hands-off data center expansion. They will walk through the original design, the failure modes that forced change, and the deliberate tradeoffs they made. The presentation will demonstrate how they used Temporal to automate previously manual workflows, removed redundant dependencies, and ultimately deprecated PgBouncer in favor of a homegrown Postgres proxy.

This practical, experience-driven talk about simplifying Postgres at scale will show how Datadog used automation to tame complexity, AI to detangle existing workloads, and built database architectures that can grow without becoming an operator's nightmare.

Speaker: Fabiana Scala
I am a Staff Engineer at Datadog, I like to focus on providing platform solutions for the product teams to focus on developing their business logic. I am also very active in mentoring other women engineers at Datadog and love to share my knowledge with my peers. Outside of work you can find me playing board games, tennis or traveling.