How Itaú Improved Payment Reliability When Every Transaction Matters
Itaú’s financial authorizer processes a massive share of Brazil’s instant payment system (PIX), operating under strict sub-500ms SLAs where every millisecond impacts revenue and customer trust. In this talk, Edinei Piovesan, Principal Software Engineer, will share how Itaú increased availability from below 99% to more than 99.95% by combining observability with operational discipline. At the center are “OpsMeetings”: automated weekly SLO reports delivered to every team, with enforced accountability through live analysis of burn rates, latency spikes, and error budgets. Edinei explains how this improved incident response while enabling teams to challenge assumptions, uncover hidden latency, and bring previously “untouchable” systems out of the black box. Learn how Itaú uses “Ask Bits” to standardize “war room prompts” and quickly answer availability and impact questions during incidents, while validating whether deployments reduce latency and SLO violations.