As Andreas Garbner mentioned here:
You can create exception count measure based on this exception. You have to open system profile, measures and create new measure there.
After this you have to set threshold there for warning and severe. This will represent how many exceptions should alert you through incident.
Then create incident rule based on this measure. You can pick there amount of time that will be considered with thresholds. For example severe incident when during 10 second timeframe we have 10 jdbcconnection exceptions.