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

Company: Finastra

Industry: Financial Services

Finastra is one of the world’s leading financial software companies, serving over 8,000 customers globally. Its products span corporate and retail banking, payments, lending, treasury, and capital markets. With its open platform, FusionFabric.cloud (FFDC), Finastra empowers clients and fintechs to innovate and collaborate across the financial ecosystem by securely exposing core banking capabilities through APIs.

Challenge

Finastra’s FusionFabric.cloud platform processes millions of API requests daily across a large, distributed Kubernetes environment. But the platform team faced intermittent connection timeouts and build failures, roughly 20% of builds were failing that were difficult to reproduce or diagnose. Despite having implemented traditional observability pillars (metrics, logs, traces), adoption across development teams was low, and developers lacked visibility into how microservices communicated and where latency or timeouts occurred.

Solution

Finastra’s platform team turned to Grafana Cloud and open source eBPF-based tooling to achieve zero-code observability across their environments.

• Grafana Beyla: Enabled automatic tracing and RED metrics collection across all services without code changes.
• Grafana Cloud dashboards: Provided unified visibility across Java, Go, and Node.js applications, creating a shared observability language for all teams.
• Grafana Cloud Application Observability : Delivered auto-instrumentation of all microservices, giving teams instant distributed traces and request flow visibility.
• Grafana Tempo: Enabled precise root cause analysis of timeouts and blocked threads, mapping service interactions and pinpointing faulty code methods.
• Grafana Alloy: Unified metrics, logs, traces, and profiling into a single lightweight agent, simplifying configuration and reducing operational overhead.

“We learned that we can do zero-code observability, and that is extremely powerful. Suddenly you have the same common language spoken by different teams in different geographies and it costs them absolutely zero to do.”

– Vladimir Babichev, Platform Architect

Impact

After deploying Grafana’s open observability cloud, Finastra’s teams rapidly identified the root causes of performance degradation and stabilized their environments.

• Reduced build failures from 20% to near zero, unblocking multiple product teams.
• Pinpointed critical timeout issues in code, enabling fixes with only minor configuration changes.
• Prevented potential production outages, improving overall platform resilience and stability.
• Lowered memory and infrastructure costs by consolidating observability agents with Grafana Alloy.
• Accelerated collaboration between globally distributed teams by providing standardized dashboards and metrics.

“We empowered our teams. There’s no more finger pointing. Instead of being problem owners, we became solution providers. And that was very powerful for us.”

– Vladimir Babichev, Platform Architect