（大佬求救，毕设相关） Prometheus、Granafa的Cluster-Overview以及Dashboard的p99延时不同

histogram_quantile(0.99, sum(rate(tidb_server_handle_query_duration_seconds_bucket{k8s_cluster=“”, tidb_cluster=~“tidb-cluster-basic”, instance=~“basic-tidb-.*”,sql_type !=“internal”}[10m])) by (le, instance)) * 1000

请问是这样写嘛，看起来貌似可行？