Application Gateway Re-IP Challenges

If you have worked with Application Gateways in Azure you’ll already know that they can be straight forward — unless you want to change the configuration. Recently I was working with a client that had to re-ip their systems as part of an integration. The result was that when we were not able to re-ip the app gateways and the virtual machines behind them in the backend pools did not update correctly even though the vm’s were bound directly and not by IP address.

We found a way to work-around this by creating the app gateways in parallel with the new IP addresses and then binding the vm’s in the backend pools. To my surprise, the backend pools also had challenges on change day.

For some reason the targets in the backend pool are not registering now after making the IP address change on the servers.

Checking in with the Backend Health blade showed that there was no communication to the servers and that health has failed. Not good.

The fix to the backend pools ended up being a simple as the old on/off adage.

We dropped the backend pool configuration, saved the changes (10 minutes and some time pondering ones existence and that I was grateful for the challenge that was presented to me).

Next we added the vm’s back into the pool, saved again, and crossed our fingers…

It worked! Dropping and re-adding the backend pool fixed the issue! Awesome!