I share the story of a frustrating debugging journey with a Rust application that was migrated to Kubernetes. Everything seemed fine until a suspected memory leak caused our application to scale uncontrollably. I dove deep into tracing memory usage with eBPF, heaptrack, and jemalloc profiling tools. The real twist came when I discovered that the memory leak was a phantom menace - a misinterpretation of Kubernetes memory metrics rather than a genuine issue. This post details the steps I took, ...
