Is the caching implementation (when I use

use_cache

as described here) a distributed cache in a multi-node typesense setup? Or are query results only cached by the node that served it?

Local to every node.