False alerts being generated

Such alerts usually show up when the pool member fails to respond to health monitors. You can confirm this by checking the logs. There are times when the pool member may not respond to a health check because it is busy processing other traffic. In that case, changing the default timeout and interval on the health-check monitors would help.

You can increase the timeout & interval within the monitor if you want to reduce such alerts. Timeout is usually 16s and Interval is 5s. You can change it to 10s Interval and 31s Timeout. Timeout needs to be [(3*Interval) + 1] as a recommended best practice.