Best Practices and Interview Q&A

Operational Best Practices

• Set maxmemory and an eviction policy before load.
• Enable AOF with everysec fsync for durability.
• Bind Redis to private interfaces only.
• Move backup snapshots off the host.
• Watch latency and memory with INFO and monitoring.

Design Best Practices

• Keep large payloads out of Redis; store references.
• Use TTLs on every cache and session key.
• Prefers hashes over many tiny key/value strings.
• Design stable keys with a namespace like app:users:42.
• Invalidate on writes to keep caches correct.

Interview Q&A

Q: Why is Redis so fast?

A: Data lives in RAM and the single-threaded command loop avoids locks, executing simple operations in microseconds.

Q: Redis vs Memcached?

A: Redis offers data types, persistence, Pub/Sub and clustering; Memcached is a simpler distributed cache with lower overhead for pure key/value.

Q: How do you ensure a cache stays fresh?

A: Use a TTL for expiry and write-through invalidation - evict or update the key whenever the underlying data changes.

Q: How do you recover a lost primary?

A: Redis Sentinel promotes a fresh replica automatically and routes clients to it; clusters fail over at the slot level.

Key Points

• Memory, TTL and eviction bound the cache.
• Type choice directly shapes performance.
• Persistence plus replicas guard against loss.
• Sentinel and Cluster scale availability.