Tailscale exit-node failover: automatic switchover when primary VPS drops

Running Tailscale as an exit node for a small homelab setup. Primary exit node is a Hetzner VPS in Nürnberg, backup is a local Raspberry Pi. Problem: When the VPS goes down (maintenance, network issue), connected clients don't automatically failover to the secondary exit node. Tailscale's exit-node selection is manual — you set `--exit-node=<IP>` or pick from the admin panel, but there's no built-in health-check + auto-switch mechanism. Current workaround: a cron script on each client that pinges 1.1.1.1 through the exit node, and if it fails, runs `tailscale up --exit-node=raspberrypi`. This is clunky and causes a 30-60 second connectivity gap. Questions: 1. Has anyone built a proper exit-node health monitor using Tailscale's status API (`tailscale status --json`)? 2. Would DERP fallback be a better approach than exit-node switching for this scenario? 3. Any experience with policy-based routing + keepalived on top of Tailscale interfaces? Infrastructure: 4 clients (2 Linux, 1 macOS, 1 Windows), all on Tailscale 1.64+. Primary exit node has 100 Mbps uplink, backup is 20 Mbps on residential fiber.