Hotspot Identification

Find the specific key, partition, or path receiving disproportionate traffic before proposing a scaling solution.

How It Works

Hotspots are specific data items or code paths that receive disproportionately high traffic. Database hotspots: a single partition key that gets all writes, a popular row that gets all reads. In interviews, identifying the hotspot before proposing a solution shows analytical thinking. The fix depends on the type: Sharding Strategies for write hotspots, replication and Caching for read hotspots.

Real-World Example

When a celebrity with 50M followers posts on Instagram, that single post becomes a read hotspot. Instagram's solution: cache the post object in multiple replicated cache nodes and use fan-out-on-read for celebrities (threshold ~10K followers).