Last week I created an azure LB routing HTTPS traffic on port 443 to an alternate port on the back-end web servers. The health probe for the LB rule was using the same alternative port. Sometime later in the week I experienced an issue with each of the web servers returning 500 errors on the alt HTTPS port. The health probe rightly removed those servers from the back-end pool. However it appears the servers were never added back into the pool once the error was resolved. The only way I could fix the issue was to open a new port on the web servers - with a new health probe and load balancing rule on that port.

When does The azure health probe add the servers back into the back-end pool? I would expect it to do so after the health probe returns 200 once or twice. however this is opaque from the health probe configuration.

Also, is there any way to gain additional configuration for the health probe state of the back-end servers such as return to normal state and flapping states?


TCP and HTTP probes are considered healthy and mark the role instance as healthy when:

  • The load balancer gets a positive probe the first time the VM boots.

  • The number SuccessFailCount (described earlier) defines the value of successful probes that are required to mark the role instance as healthy. If a role instance was removed, the number of successful, successive probes must equal or exceed the value of SuccessFailCount to mark the role instance as running.

More information please refer to this link.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.