Title

AWS re:Invent 2022 - A real-world resilience evolution in the cloud framework (ARC309)

Summary

Anderson Motta from Itaú and Gus Santana and Robert Fuente from AWS presented their collaborative resilience program.
Itaú Unibanco, a 98-year-old financial institution from Brazil, is the largest in Latin America with operations in 20 countries and a strong digital customer base.
The journey to the cloud began in 2016 with private cloud frameworks and data centers, evolving to a public cloud-first mandate in 2020.
Itaú has migrated 45% of its applications to AWS, aiming for 60% by the end of 2023.
The Resilience Evolution Program was created to address high-impact events affecting customers, with AWS working backwards to meet Itaú's needs.
The program established five pillars: monitoring, testing (chaos engineering), understanding dependencies, architectural enhancements, and governance.
A "mechanisms" pillar was added to scale the lessons learned across the bank.
The program led to a 75% reduction in high-impact events and a significant reduction in Mean Time to Recovery (MTTR).
AWS provided a Single Threaded Leader (STL) for the program, daily stand-up meetings, and a focus on communication.
Key activities included chaos engineering, critical dependencies analysis, blue-green deployment strategies, and observability improvements.
AWS Incident Detection and Response (IDR) was introduced to enhance support for critical workloads.

Itaú's digital transformation required decoupling monolithic applications into microservices and adopting a data-driven approach with observability.
The Resilience Evolution Program is an example of AWS's customer-centric approach, creating a tailored solution for Itaú's specific challenges.
The program's success was measured not only by the reduction in high-impact events but also by the establishment of a sustainable model for future application development.
The STL role was crucial in ensuring the program's focus and accountability, demonstrating the importance of dedicated leadership in large-scale projects.
The use of chaos engineering and dependency mapping highlights the proactive approach to resilience, preparing for potential failures rather than reacting to them.
The creation of executive dashboards and traffic light systems (farol) for application health visibility indicates a high level of maturity in operational monitoring.
The introduction of AWS IDR shows a commitment to continuous improvement and the value of feedback loops between AWS and its customers.
The program's emphasis on governance and executive sponsorship underscores the importance of aligning technical initiatives with organizational support and change management.