# Sql Query Time-Out Case Study Background

+ Cluster configuration: 3 PD nodes, 19 TiDB nodes, and 24 TiKV nodes
+ Observation: At around 17:12, the application reported that a large number of queries timed out, and then recover after 1~2 minutes. The DBA received a slow query alert. The slow query SQL statement is**PointGet** point query, and the delay reaches seconds

------------------------------------------------------------

Case background: This is an Internet customer, with a large cluster configuration, 3 PD nodes, 19 TiDB nodes, and 24 TiKV nodes.

Phenomenon: At around 17:12, there was a business feedback: the query timed out during this period, the performance problem lasted for 1 to 2 minutes, and then automatically recover. Explain that the query was not always slow, it happened occasionally. Then the DBA received a warning about slow query, which SQL was slow in the slow query log, and found that the slower SQL was all point query. It is a very common phenomenon to know that the key goes to a TiKV to check the value in a region with a delay of seconds.