The load balancer can probe the health of the various server instances. When a probe fails to respond, the load balancer stops sending new connections to the unhealthy instances. Three types of probes are currently supported:

• Guest agent probe (on PaaS VMs only): The load balancer utilizes the Azure guest agent inside the virtual machine. The guest agent listens and responds with an HTTP 200 response only when the instance is ready and healthy. If the agent fails to respond with an HTTP 200 response, the load balancer marks the instance as unresponsive and stops sending traffic to that instance. The load balancer will continue to ping the instance until it responds again or the instance is removed from the load balancer set. Attention: If you are running a website, the code is typically running the process w3wp.exe. This processes are not monitored by the guest agent so the load balancer will never informed when the instance fails.
• HTTP custom probe: This probe overrides the default (guest agent) probe. You can use it to create your own custom logic for your application to determine the health of the role instance. As an example, the HTTP probe could login to a page and changes some values as a result of the logic. If the when fine without issues, the probe is good, otherwise it will generate an alert. The load balancer will regularly probe your endpoint (every 15 seconds, by default).
• TCP custom probe: This probe relies on successful TCP session establishment to a defined probe port.