You should monitor the following metrics on the load balancer.
- Healthy host count. The number of healthy instances. An instance is considered healthy if it meets the healthy threshold configured for the health checks.
- Unhealthy host count. The number of unhealthy instances. An instance is considered unhealthy if it exceeds the unhealthy threshold configured for the health checks. The count should not be greater than zero unless the host is under maintenance.
- Request Count. The number of completed requests that were received and routed to the registered instances.
- Latency. The time elapsed, in seconds, after the request leaves the load balancer until a response is received. The number should not be greater than zero unless the host is under maintenance.
- Backend Connection Errors. The number of connections that were not successfully established between the load balancer and the registered instances. The number should not be greater than zero unless the host is under maintenance.
- Surge Queue Length. The total number of requests that are pending submission to a registered instance. The number should not be greater than zero unless the host is under maintenance.
- Spillover Count. The total number of requests that were rejected because the queue was full. The number should not be greater than zero unless the host is under maintenance.