The percent of prunes from the query cache. Query cache pruning occurs when the query cache is full and the least recently used query is pruned. The query cache size is controlled using the global variable “query_cache_size”. The monitor threshold is analyzed against an average of values returned in polls to the agent where the Number of Samples parameter controls the number of samples included in the calculation of the average.
An unhealthy state indicates the percentage of query cache prunes is high. The query cache might not have enough storage for recently used queries. The query cache limit for individual queries may be too high causing more pruning.
To address high query prune percent increase the query_cache_size to allow additional queries to be stored. Decreasing query_cache_limit can allow more queries to fill up the query cache before a prune.