You can broadly categorise goalkeeping ability down to shot stopping, command of area (claiming crosses, sweeping actions etc), and distribution.

Most of the modern world class keepers you see are very good in all of these (think alisson or neuer).

Shot stopping you can kind of simplify down to xGsave overperformance over an extended period of time, small sample sizes aren’t really accurate for this.

Distribution you can also use progressive passing metrics, long ball success rate, things like that. There’s also an element of eye test in terms of how comfortable they look on the ball e.g. do they manipulate it well when under pressure (ederson) or hoof it out of play when pressed? (De Gea).

Command of area is much more difficult to quantify because it’s a preventative action and doesn’t show in data as easily, for example neuer sweeping and preventing a 1v1 won’t show in the xG data, but it has a massive effect on the game. Or nick pope claiming a cross from a corner that would have found the head of an attacker is hard to quantify.There might be some metrics for this that exist but they don’t come up often in the mainstream.

Being very good at 2 of those 3 things make a fantastic keeper, all 3 and they’re probably elite or have very high potential.

A team like Luton won’t require their keeper to be good at distribution, but Bayern most certainly will, so they won’t really shop in the same market for goalkeepers, on the other hand ederson might not be top tier for shot stopping, but his build up is so important for city that it doesn’t really matter. They also don’t face a huge number of high quality shots game to game.

That’s the way I perceive goalkeeping ability anyway.