Answer
If every server's check fails when the shared database has a hiccup, the load balancer marks all servers unhealthy at once and takes the whole service down, even for requests that did not need the database. Keep the routing check about the server itself, and handle dependency failures with degradation instead.