Answer: Cache used percentage, Cache hit percentage

## Detailed Explanation When diagnosing query performance issues related to columnstore segments in Azure Synapse Analytics SQL pool, the key metrics to monitor are **Cache used percentage** and **Cache hit percentage**. ### Why B (Cache used percentage) and D (Cache hit percentage) are correct: **Cache used percentage** indicates how much of the available columnstore cache is currently being utilized. When this metric approaches 100%, it suggests that the cache is fully utilized, which can lead to performance degradation as new queries may need to fetch data from the underlying storage rather than from the faster cache. **Cache hit percentage** measures the percentage of queries that successfully retrieve data from the columnstore cache rather than having to read from the underlying storage. A low cache hit percentage (typically below 90-95%) indicates that queries are frequently reading from storage, which is significantly slower than reading from cache, directly contributing to the reported performance issues. ### Why the other options are incorrect: **A (Snapshot Storage Size)**: This metric relates to the size of database snapshots and their storage consumption. While important for storage capacity planning, it has no direct impact on query performance or columnstore segment efficiency. Monitoring snapshot size won't help diagnose columnstore-related performance issues. **C (DWU Limit)**: This represents the Data Warehouse Unit limit configured for the SQL pool. While DWU settings affect overall resource allocation and performance, this metric doesn't provide specific insights into columnstore segment performance or cache utilization patterns that are causing the reported query slowdowns. ### Best Practice Context: In Azure Synapse Analytics, columnstore indexes are fundamental for analytical query performance. The columnstore cache stores frequently accessed column segments in memory for faster retrieval. When queries take longer than expected and the issue is traced to columnstore segments, monitoring cache utilization patterns is the most direct approach to identify whether the performance degradation stems from insufficient cache capacity (high cache used percentage) or inefficient cache usage (low cache hit percentage).

Author: LeetQuiz Editorial Team