# Coprocessor（B）

+ Grafana —> TikV-Details —> Coprocessor Detail —> Wait duration
  + Sum of the time the request was scheduled+get the Snapshot+ build the Handler
  + The request is scheduled, and the time to get the snapshot and build the Handler is very short, not a bottleneck

![left](./student-guide/diagram/27-Coprocessor-02-01.png)
------------------------------------------------------------
- Grafana —> TikV-Details —> Coprocessor Detail —> Wait duration
Wait duration: the sum of the time the request was scheduled+get the Snapshot+ build the Handler (the delay for max/.99/.95 is included in the monitoring)

  - The delay when the request is scheduled does not involve IO or network, only the load on the CPU. If the CPU load is not high, the scheduling time should be negligible
  - The time to build the Handler is negligible even if the CPU load is not high
  - If the snapshot acquisition time is too long, you need to troubleshoot the Storage async-snapshot duration monitoring metric

In summary: You should first look at the CPU load. If the CPU load is high, the bottleneck is in the CPU, otherwise further troubleshoot the cause of slow snapshot acquisition
The demand schedule, and the time to get snapshots and build the Handler is very short, not a bottleneck