A 503 Service Unavailable from a Vultr Load Balancer means the load balancer cannot forward requests to any healthy backend instance. This happens if no instances are attached, or if all attached instances are failing health checks. To fix this:
- Ensure all instances attached to the Vultr Load Balancer are in a Running state. Stopped or terminated instances will trigger 503s.
- Health checks determine if the load balancer considers an instance available. Verify that the health check protocol, port, and path match your application. For HTTP checks, make sure your service responds with a
2xx or 3xx status on the configured path within the Response timeout window.
- Make sure the application on each backend instance is running, listening on the correct port, and can serve requests from the load balancer’s IP range.
- Confirm that Firewalls allow incoming traffic from the load balancer. If using private networking, ensure the instances are in the same VPC as the Vultr Load Balancer.
- If the PROXY Protocol is enabled on the Vultr Load Balancer, the backend must be able to parse it. Misconfiguration will make the instance appear unhealthy.
Following these steps ensures your Vultr Load Balancer routes traffic correctly and eliminates 503 errors caused by unavailable or misconfigured backends.