Tokio Gives Progress, Not Ordering: Scheduling 1M Tasks

# Summary The author describes how their Rust service using Tokio spawned ~1 million short-lived tasks during event bursts (1000 events × ~1000 user tokens each), and noticed that tasks from earlier events were being scheduled and executed much later than expected, despite completing within the required time. This investigation into task scheduling behavior was a precursor to discovering that their memory issues were actually caused by the memory allocator rather than the application code or task management.