An interactive explainer
A cache serves 99 of every 100 reads. The database behind it has never seen more than a trickle. One popular key expires and that database gets every read at once. Then the retries start, and they arrive because it got slow.
Caching · Cache stampede · Retry storms
Systems · Advanced · 11 min · Why one expired key takes down a database that was 20% busy — an interactive explainer on ViLearn, with a simulation and a knowledge check. All explainers.
Related: What a queue really does when you are overloaded · How a cluster decides which server holds your data · How a service decides who gets in