Retry Storm
A failure-amplification pattern in which many clients retry a slow or failing dependency at the same time, adding load precisely while the dependency is least able to absorb it.
When a dependency is near saturation, immediately repeating failed calls can multiply incoming work instead of improving success. Similar timeout and backoff settings can also synchronize clients, creating repeated traffic waves that push a recovering service back into overload.
Exponential backoff, jitter, retry budgets, and admission control therefore belong to the same control loop. Retries should be reserved for transient failures and constrained by the end-to-end deadline. Safe retry design in critical systems covers the related idempotency and error-classification issues.