Is there a way, for a set of routes, to apply rate-limiting or lower priority or just confine them to a fixed number of threads so that they don't overrun the server? We're using jetty...

if they've made it into your routes, jetty is already handling them you may want to consider a rate-limiting load balancer in front of your service (e.g. API Gateway)

in-app rate limiting makes sense IMO, if the resource being rate-limited isn't jetty itself e.g. if some route guards some special resource, maybe you want to rate-limit that specific resource

I should probably look into jetty specifically, and see if it can do basic routing and strategies like this before handing over to pedestal itself. Eg a different queuing/buffering strategy.

Unfortunately can't use api gateway yet. Or can I? Not sure if it can route to an ALB…

you can use a rate-limiting interceptor

rate information could be local or shared across instances (e.g. using a db) depending on requirements