Probabilistic Early Expiration

Probabilistic early expiration refreshes some cache entries before their hard expiry using a randomized decision. Refresh becomes more likely as expiration approaches, often taking the cost of regenerating the value into account.

The goal is to spread regeneration work instead of having every reader miss at the same instant. It still requires a strategy for duplicate refreshes and failures, and the chosen probability controls the tradeoff between freshness and extra computation.