This release ships a self-healing stable-egress gateway for the hosted Tuist stack, removing the single-node dependency that caused the recent 503 outage.
- The server now egresses through a redundant pair of nodes. A new controller automatically moves the Hetzner Floating IP and the active gateway label to a healthy node when the current one fails, and it refuses to use an IP outside your reserved allowlist.
- Failover is seamless when the standby node has already served as a gateway. The first time a brand-new node takes over, convergence takes about 40 seconds, then heals automatically.
- The rollout is designed to be blip-free: the controller adopts the existing gateway node rather than moving the IP immediately, then migrates onto the dedicated pool on the next natural node replacement.
- The earlier Keygen boot hotfix is reverted, because the root cause was egress black-holing, not license validation.