You can check by the following methods:

1. Ensure that the back-end ECS can provide services normally. Does the back-end server open the corresponding ports and provide access normally? The port must be consistent with the back-end port configured in the load balancing listening configuration;
2. Log in to the console to check whether the listening configuration is correct. Check whether the front and rear ports in the monitoring are correct? The check service returns a status code. If the status code of the health check does not match, the load balancing instance will reject the back-end server;
3. Ensure that the load average health check network address is in the firewall white list of the backend ECS, Load balancing health check network address ；
4. Check whether the response time of the backend server to the request exceeds the check failure time window set by the health check;
5. Network check. Telnet<Public IP address of load balancing><Port of external monitoring of load balancing>
6. Check the backend server separately. Access curl - I from other ECS<ECS intranet IP address><Port>
7. Check whether the backend ECS resources have a high load or other reasons that lead to a long response time for external service delivery
8. It is recommended to use HTML static file pages as the URI of health check, and it is not recommended to use PHP and other dynamic script language pages, because the dynamic script may be removed from the load balancing cluster when it changes, resulting in service unavailability.