# Objectives

+ Differentiate performance issues between TiDB Server and TiKV Server
+ Differentiate TiKV write latency from external causes such as thread pool bottlenecks or disk performance jitter
+ Monitor the impact of disk capacity on PD scheduling and performance

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

This lesson shares three cases:
First, when you encounter a performance problem, how to distinguish between a TiDB Server problem or a TiKV problem.
Second, if lockdown is a problem with TiKV, what exactly is wrong with TiKV? Is it a thread problem, an IO problem, or a hardware problem?
Third, if the PD has problems with response and scheduling, it has an impact on performance. So how do you monitor lockdown issues?