> Am I right in understanding that this is the tricky balance you need to strike? And shooting for Pe=Pa is where it happens to be optimized for thrust?

The tradeoff is complicated.

First stage nozzles are always compromises. A smaller nozzle is a better match at takeoff and it lets you fit more nozzles underneath the base of the rocket. A larger nozzle is better later in the first stage burn as it's a better match to the lower ambient pressure. However, if it's too big there will be significant flow separation at sea level and that can break the nozzle.

And it gets more complicated with design choices. Rockets like the Falcon 9 stage fairly early, so they have nozzles that are decent compromises between sea level and vacuum. Rockets like the Atlas V stage much higher and therefore the first stage spends more time in vacuum. And rockets like the shuttle run their engines all the way to orbit, so vacuum performance is far more important.

Second stage engines are generally easier; you generally just put on the largest nozzle that will comfortably fit.