Does a coprocessor graph like this indicate poor performance?

Translated
1

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: coprocessor这样的图算不算性能差？

| username: vincentLi

[TiDB Usage Environment] Production Environment
[TiDB Version] 6.5.3
[Encountered Problem: Phenomenon and Impact]
Many slow queries have long coprocessor times. Due to watermark issues, here is the “coprocessor request duration” graph from Grafana, sorted by max from high to low.
max current
select-99% 40.48648ms 637.9200us
select-95% 38.26865ms 294.6717us
index-99% 30.7712ms 2.5516ms
select-avg 12.80510ms 135.073us
index-95% 5.06892ms 2.49193ms
index-avg 2.27903ms 1.53678ms
If this data is poor, how can we further analyze whether it is due to insufficient nodes, CPU, memory, or IO?

2
| username: 小龙虾爱大龙虾 | Original post link

This metric cannot analyze the issues you mentioned, such as insufficient nodes, CPU, memory, or IO.

3
| username: forever | Original post link

Take a look at the overall resource situation of the cluster.

4
| username: zhaokede | Original post link

I can’t tell. Could you provide the execution plan and related server configuration?