The statistics and actual queries of monitoring are inaccurate, what is the reason?

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

Original topic: 监控的统计和实际查询不准,啥原因

| username: 路在何chu

[TiDB Usage Environment] Production environment 4013
[Reproduction Path] What operations were performed to cause the issue
Monitoring shows the current select is over 2000

[Encountered Issue: Issue Phenomenon and Impact]
However, the database query is only a few hundred. The internal SQL of the database is not counted, but the internal SQL shouldn’t be that many, right? The image below shows the select volume over 20 minutes, cleared every 30 minutes.

| username: DBAER | Original post link

Shouldn’t multiple TiDB instances be aggregated?

| username: 路在何chu | Original post link

Just one, not multiple.

| username: 路在何chu | Original post link

Oh, I forgot. I need to check CLUSTER_STATEMENTS_SUMMARY.

| username: tidb菜鸟一只 | Original post link

Yes, the entire cluster needs to query tables with the cluster prefix.