I've noticed recently that our resting CPU on our Aurora RDS cluster appears to be about 80%. I'm unsure when this started, as we don't have alarms/metrics in place (something I plan to rectify now).
It's a 2 node cluster db.t2.medium. Normally with a MySQL/MariaDB instance I'd troubleshoot on the actual machine, however the trade off of cloud services is that I can't really do this.
We have a script that normalises some data every 10 minutes, and the spikes I see there reflect the 10-15% jump, which I'd expect. However we appear to be constantly resting around the 70-80% mark. The update, insert and delete throughput per second never exceeds 4/s, so I can't imagine other operations are affecting it.
Why would this be, and how can this be troubleshot?