An Azure service that provides a cloud content delivery network with threat protection.
the apex problem is a DNS limitation, rather than one specific to Application Gateway for Containers. Your Traffic Manager approach is conceptually valid, but the nested Traffic Manager design is probably unnecessarily complicated and introduces another DNS-resolution layer. You might consider a simpler approach :
example.com → Azure DNS alias (A/AAAA) → Traffic Manager profile → External endpoint → Application Gateway for Containers FQDN
In other words, use a single Traffic Manager profile whose external endpoint target is the Application Gateway for Containers frontend FQDN. Traffic Manager is capable of using an FQDN as an external endpoint target, so you should not need Profile 1 → nested Profile 2 just to get from the apex to the Application Gateway for Containers hostname.
The Azure DNS portion would be an alias A record at the zone apex targeting the Traffic Manager profile. Traffic Manager then resolves its external endpoint, for example xxxx.<region>.alb.azure.com, and returns the resulting address to the client. Because the Application Gateway for Containers FQDN remains the Traffic Manager endpoint target, changes to the underlying Azure-managed IP address do not require you to modify DNS manually.
I would avoid the current Profile 1 → nested Profile 2 → external endpoint chain unless you have a specific requirement for two independent Traffic Manager policies. The fact that substituting the currently resolved IP address makes your configuration work seems to indicate that the failure is occurring while Traffic Manager is resolving or processing the FQDN used as the external endpoint, rather than at Azure DNS or at Application Gateway for Containers itself.
Note that there is a distinction between DNS resolution and HTTP/S routing here. Traffic Manager is DNS-based. It does not proxy the HTTP request to Application Gateway for Containers. Once the client receives the IP address returned through the Traffic Manager resolution process, the client connects directly to that IP. Effectively your Application Gateway for Containers configuration must still have example.com configured as an appropriate hostname/listener and the corresponding TLS certificate must cover example.com.
If a single Traffic Manager profile with an external endpoint whose target is the Application Gateway for Containers FQDN still exhibits intermittent regional failures, I would still advise against hard-coding the current IP. The IP behind the Azure-managed frontend is not a stable contract, so doing that defeats the purpose of using the managed FQDN.
For a production deployment, I would use:
example.com → Azure DNS alias A/AAAA record → Traffic Manager profile → External endpoint with target <Application-Gateway-for-Containers-FQDN> → Application Gateway for Containers → AKS
and separately:
www.example.com → CNAME → <Application-Gateway-for-Containers-FQDN>
If you need Traffic Manager's health-based routing, priority/weighted routing, or multiple Application Gateway for Containers frontends, you can still use those capabilities, but I would keep the Traffic Manager hierarchy as shallow as possible.
If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.
hth
Marcin