This is fun, but I'm confused by the architecture. Duckdb is based on one-off queries that can scale momentarily and then disappear, but this seems to run on k8

This is fun, but I'm confused by the architecture. Duckdb is based on one-off queries that can scale momentarily and then disappear, but this seems to run on k8s and maintain a persistent distributed worker pool. This pool lacks many of the features of a distributed cluster such as recovery, quorum, and storage state management, and queries run through a single server. What happens when a node goes down? Does it give up, replan, or just hang? How does it divide up resources between multiple requests? Can it distribute joins and other intermediate operators? I have a soft spot in my heart for duckdb, but its uniqueness is in avoiding the large-scale clustering that other engines already do reasonably well.