Hi - I’m using pmm to monitor db.r5.xlarge RDS aurora mysql instances with enhanced monitoring turned on and we’ve found that the Percona was pegging the CPU while gathering the metrics . Is this expected or is there a setting that needs to be adjusted as such ?
1 Like
Hi - Can anyone please confirm this or suggest on this
1 Like
That’s not at all expected. What is your metrics resolution and how active is this database? QAN can have an impact if you’ve enabled full query samples but shouldn’t peg it by any stretch. Are you using slow log or performance schema? We typically see pmm consume somewhere between 2 and 5% of cpu on clients and we’ve done quite a bit of performance optimization for RDS and Aurora. Anything special about this instance?
1 Like