With eBPF and OpenTelemetry, Finastra turned a potential internal DoS into a new observability architecture built for long-term resilience

What happens when a mysterious “Connection Read Timeout” starts randomly taking down microservices, and no one knows why?

In this talk, Finastra Platform Architect Vladimir Babichev and Expert Software Engineer Maciej Maciejewski will recount how the fintech company’s overnight deployment of Grafana Alloy (a distribution of the OpenTelemetry Collector), Grafana Beyla (eBPF-based autoinstrumentation), and the OpenTelemetry Operator not only uncovered the root cause of a long-standing instability and prevented a full-blown internal DoS. It also led to future-proofing the organization’s observability architecture.

In this case, logs and traces weren’t enough – until the team found the right ones, with the help of eBPF and OpenTelemetry. The team will share the surprising insights and patterns discovered along the way, and the lessons learned that the team applied to redesign Finastra’s observability architecture for long-term resilience.