Configuring load balancer health checks for Cloud Foundry routers

Page last updated:

You can configure load balancer health checks for Cloud Foundry routers so requests go only to healthy router instances.

In environments that require high availability, you must configure your redundant load balancer to forward traffic directly to the Cloud Foundry (Cloud Foundry) routers. In environments that do not require high availability, you can skip the load balancer and configure DNS to resolve the Cloud Foundry domains directly to a single instance of a router.

Add health check endpoints for routers

You can configure your load balancer to use either HTTP or HTTPS for checking the health of gorouter and tcp_router processes.

To deactivate the HTTP health check endpoints, set the following properties to false: - The router.status.enable_nontls_health_checks property on the gorouter job - The tcp_router.enable_nontls_health_checks property on the tcp_router job

Using HTTPS health check endpoints

To configure your load balancer to use HTTPS health check endpoints, add the IP addresses of all router instances, along with their corresponding port and path:

• Gorouter (HTTP router): https://GOROUTER_IP:8443/health
• TCP router: https://TCP_ROUTER_IP:443/health

Using HTTP health check endpoints

To configure your load balancer to use HTTP health check endpoints, add the IP addresses of all router instances, along with their corresponding port and path:

• Gorouter (HTTP router): http://GOROUTER_IP:8080/health
• TCP router: http://TCP_ROUTER_IP:80/health

The preceding configuration assumes the default health check ports for the Cloud Foundry routers. To modify these ports, see the following sections.

Set the Gorouter Health Check Port

You can set the health check port for Gorouter in the cf-deployment manifest using the router.status.port property. The value of this property default setting is 8080.

Set the TCP Router Health Check Port

You can set the health check port for the TCP router in the routing-release manifest using the tcp_router.health_check_port property. The value of this property default setting is 80.

Set the healthy and unhealthy threshold properties for the Gorouter

To maintain high availability during upgrades to the Gorouter, each router is upgraded on a rolling basis. During upgrade of a highly available environment with multiple routers, each router is shut down, upgraded, and restarted before the next router is upgraded. This ensures that any pending HTTP requests passed to the Gorouter are handled correctly.

Cloud Foundry uses the following properties:

router.drain_wait: Specifies, in seconds, the unhealthy threshold that determines when the Gorouter stops accepting connections and the process gracefully stops. During this period, the Gorouter continues to serve HTTP requests and the health check endpoint returns 503 errors.

router.load\_balancer\_healthy\_threshold: Specifies the amount of time, in seconds, that the load balancer waits until it declares the Gorouter instance started. This enables the load balancer time to register the instance as healthy.

The following table describes the behavior of the load balancer health checks when a router shuts down and is restarted.