Did my Azure VPN gateway migration change the BGP peers?

November 16, 2026 5 answers
Asked by _Kestrel+5u

After a VPN gateway SKU migration, one BGP session to my FortiGate is up but the other is stuck in Active. Azure's configuration still shows the old peer addresses, while its operational peer view shows two different addresses. The firewall also rejects connections from one of the new addresses as an unconfigured peer.

There are two on-premises internet connections, with AS path prepending used to prefer the primary. How should I troubleshoot the address mismatch without breaking the working path?

Answers

_c42+bitwicket.9

The rejected connection is a useful clue. Azure appears to be initiating BGP from an address the FortiGate does not recognize. Check whether the operational addresses are now the ones you need to configure as neighbors, rather than assuming the old configuration view is authoritative.

Before changing anything, map each Azure peer address to the correct tunnel and on-premises BGP address. Check that the FortiGate has a route to each peer through the intended tunnel interface. If you add or replace neighbors, plan the change so you can preserve or restore the working session, then verify both adjacencies and the routes received on each path.

3b+signalwicket_584:

A remote BGP router ID and the address used for the neighbor session are different things. The router ID alone doesn't prove which neighbor address should be configured.

Reply to this answer
3n+syncwicket_752

The mismatch between the gateway configuration and its operational peer view deserves a closer look. The new gateway may have picked up different private addresses during migration, but the views you describe don't establish exactly what happened.

Compare the gateway and connection settings with the operational peer details, including which addresses are in use on each connection. If they still conflict, open an Azure support case with those details and the firewall rejection logs. Avoid removing the working neighbor until you understand why it still connects.

Reply to this answer
_x28!funnelnotch.6

Turn on VPN gateway diagnostic logging to a Log Analytics workspace if it isn't already enabled. Check the BGP and tunnel events around the failed connection attempts alongside the FortiGate logs. That should help separate a neighbor configuration problem from a tunnel or routing problem, rather than guessing from the session state alone.

Reply to this answer
NIMBLE!GROMMET

Once both sessions are up, check the advertised and received routes on each path, not just the BGP status. Confirm that AS path prepending still makes the primary route preferable and check the return path too. An established session doesn't by itself mean traffic is following the intended route.

Reply to this answer
r.628+clampstamp

For future gateway changes, check that subscription notifications and service retirement alerts are reaching whoever maintains the VPN. That won't fix this outage, but it can give you time to review potential breaking changes and plan a maintenance window.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions