We believe we have identified a bug within the Azure Portal.
See steps below:
- Go to the Azure Application Gateway service
- Click on any current Application Gateway
- Go to the HTTP settings section and then click on the current config set, the port number changes to the default HTTP port of 80, instead of keeping the port that is already configured.
In our case, we were going into the configurations section to only adjust the timeout value, not the port, the port was automatically changed and this caused a temporarily outage as we did not realize until afterwards that this changed the port. It caused the backend to be marked as unhealthy, because it was not communicating over port 80. I would recommend this section to keep the current configured port when making changes to something like the timeout setting. I attached a screenshots for reference.