How One CTO Used Observability to Find a Memory Leak Saving Millions

Tech Leadership with Fexingo: Engineering Managers, CTOs, and Technical Leadership Conversations · Jun 13, 2026 · 6m · transcript

Lucas and Luna dive into a real-world story from a mid-sized SaaS company where the CTO used observability data — not intuition — to locate a subtle memory leak that was silently inflating their cloud bill by $40,000 a month. They walk through the specific signals: a gradual tail-latency creep, a steady rise in container restarts, and a single anomalous metric in their Prometheus dashboard that everyone had dismissed as noise. The hosts explain how the team built a custom alert based on RSS (residual sum of squares) to detect the leak before it impacted customers, and why this case proves…