What is the best practice for avoiding cache stampede (thundering herd) in Azure Cache for Redis when a popular key expires?
- A Use a very long TTL
- B Implement probabilistic early expiration (PER) or lock-based cache miss handling (correct answer)
- C Disable key expiry
- D Use WATCH/MULTI/EXEC transactions
Shortcut: Cache stampede occurs when many clients simultaneously re-compute a popular expired key. Probabilistic early expiration refreshes before expiry, or use a distributed lock to ensure only one client populates the cache on miss.