Why does an error occur when selecting sys.innodb_lock_waits in PMM?

I’m using pmm and when I try to look up sys.innodb_lock_waits in the panel edit to look up mysql’s lock holder and waiter information on the dashboard, I get an error saying db query error: query failed - please inspect Grafana server log for details. What could be the reason?

I am using AWS RDS and the version is mysql 8.0.32.

image

Welcome to Percona Community @wanskim,

The error occur because PMM isn’t giving you an SQL prompt but graphQL. You cannot execute select * to the metric there. Also, it isn’t storing innodb_lock_waits “table” unless you explicitly export it in some specific format (I don’t know yet).
If you want what you’re querying, execute it against database and not PMM.

Thanks,
K

thank you for your reply

This error occurs when querying a table in the DB through a data source. Other tables do not generate errors, but the above error occurs only when querying sys.innodb_lock_waits.

Even tables other than that table (information_schema.innodb_trx) are searched well.

oh! I see in the screenshot here that you’re using datasource as “mysql-1” a database. Well, the question is do you have sys schema on database? can you run this query directly on database?

Yes, tables from other schemas work fine.

As you mentioned, it’s AWS RDS mysql 8.0.32. It is most likely an issue for pmm rds exporter and AWS RDS mysql version.

Please report this as bug here https://perconadev.atlassian.net
You should provide the below details in the report, we will test and try to reproduce the same issue once you report this to https://perconadev.atlassian.net

  • AWS RDS MySQL and PMM version
  • How you are adding RDS instance to PMM monitoring.
  • Error details/ screenshot with error.