Which mitigation is recommended for hot spots by isolating hot keys to dedicated shards?

Which mitigation is recommended for hot spots by isolating hot keys to dedicated shards?

• A. Isolating hot keys to dedicated shards
• B. Increasing shard count without strategy
• C. Relying on a single shard for all hot reads
• D. Disabling caching

Answer: (A)

Hot spots in a sharded system happen when a small set of keys handles a disproportionate amount of traffic, causing some shards to saturate while others stay underused. Isolating those hot keys to dedicated shards mitigates this by moving the heavy load off the general shard pool and giving those keys their own resources. This reduces contention, frees up cache and I/O for the rest of the data, and allows targeted scaling of just the hot-key shards as demand grows. Increasing shard count without strategy may spread load but won’t guarantee isolation and can add management complexity. Relying on a single shard for all hot reads concentrates load and worsens bottlenecks. Disabling caching removes a major performance lever and doesn’t address how traffic is distributed.

Hot spots in a sharded system happen when a small set of keys handles a disproportionate amount of traffic, causing some shards to saturate while others stay underused. Isolating those hot keys to dedicated shards mitigates this by moving the heavy load off the general shard pool and giving those keys their own resources. This reduces contention, frees up cache and I/O for the rest of the data, and allows targeted scaling of just the hot-key shards as demand grows. Increasing shard count without strategy may spread load but won’t guarantee isolation and can add management complexity. Relying on a single shard for all hot reads concentrates load and worsens bottlenecks. Disabling caching removes a major performance lever and doesn’t address how traffic is distributed.