> I don’t know how much has been launched all the way out to L2 before...

There is a handy list on wikipedia: https://en.wikipedia.org/wiki/List_of_objects_at_Lagrange_po...

Your intuition is quite right, there is way less satellites parked around L2 than in low earth orbit for example.

But that doesn't really pose too big of a challenges in itself. It is of course far, both distance wise and energetically.

The main complication often mentioned is if something is wrong with the telescope it makes it very unlikely that a crew can visit it to fix it. The way for example how they repaired Hubble is unlikely to happen with Webb.

Meh, but there's nothing to hit in space. The energy required to reach a routine geosynchronous orbit is already ~85% of escape velocity. The added boost is minimally more dangerous and the trip is just empty hours. "Time" does kill spacecraft, but not often.

The Webb self-assembly is absolutely where the scary bits lurk.

Ariane 5 had some early problems, the first failure even became a case study in critical software development. But it had only had one partial failure after the 17th launch: the payload was launched to the wrong orbit because the wrong coordinates were put in the computer. A huge QA problem but not the fault of the launcher which did exactly as told.

Now, both Ariane 5 and Atlas 5 are extremely reliable, mature rockets.

Four of those were in the first 15, the last recent one was in 2018, and was a partial failure.