Definition

A hot partition (or hot shard/key) is a data partition that receives disproportionately high traffic, causing resource saturation.

Core concept

In a sharded system, if one shard/key sees most reads/writes, it becomes a bottleneck. That node may run out of CPU, memory, or I/O capacity, causing high latency or failures, while other shards are idle.

Use cases

This often happens with poorly chosen partition keys or time-based partitions. E.g., in a time-series DB, the latest time partition may be “hot” as new data arrives. Or a social network user with many connections might create a hot key.

Trade-offs

Hot partitions degrade performance and can lead to cascading failures if not mitigated (e.g. by adding more replicas or adjusting traffic).

Example

If userID 1234 is extremely popular, all requests for that user hit the same partition, overwhelming it. Techniques like splitting keys, caching, or read-replicas can alleviate this.

References

“Hot Shard/Partition: a shard’s resource saturation resulting in backlog of requests”[[34]](https://medium.com/@_sidharth_m_/how-to-handle-hot-shard-problem-7f640a5444e0 "How to Handle Hot Shard Problem? | Medium").