- Unbalanced shard allocation across nodes
- High network latency between nodes
- Frequent index updates
- Insufficient disk space on data nodes