If you perform 'Range Sharding' using a monotonically increasing Shard Key (e.g., a simple `Timestamp` or `SerialID`), what specific performance failure are you most likely to encounter?

1. A The cluster will achieve perfect, automatic balancing of all incoming data across all existing nodes.
2. B The database will reject the writes because the IDs are being generated too quickly for the storage engine to keep up.
3. C Overall system latency will improve because the database can predict where the next record will land.
4. D You will create a massive 'Write Hotspot' on the very last shard, as all new incoming traffic (by definition) belongs to the most recent range.