I have latest PMM version (today upgraded from 1.8.0 to 1.9.0) running in Amazon AWS monitoring few RDS Aurora instances via performance_schema. Server installed using AMI provided by Percona. Weird thing is that PMM monitoring itself puts unusually high load to the database servers. Over the time SELECT query count from PMM just keeps increasing. Eventually query analytics TOP is filled with queries by PMM. To illustrate constantly increasing query count here is a screenshot from cloudwatch showing SELECT query count for the period of 2 weeks on 2 of our DB instances. The drop on 04/02 was me restarting PMM server.
Anyone experience similar issues?