Hello Nick,
The incident related to Cloud SQL that was reported yesterday(September 4th,2019) is set to have been resolved for all affected instances.
I would suggest in this case, to check the number of connections to your instance as this could possibly have an effect on the latency as well as resource usage such as CPU.
If that is not the case, I would suggest to open a new private issue on
Public Issue Tracker page, and provide the project ID as well as the instance name, as this issue may require to be further investigated.