Azure Traffic Manager profile setting the endpoint status to degraded and switching traffic over to the different endpoint

Question

Hi folks,

Hope you are doing well!

We have a traffic manager profile and have created two endpoints, and both are setup to work as active and passive i.e. we manually switch the backend application services from one node to another and the endpoint should be also switched to the available endpoint (of which backend is served with the online services).

Now, for past few days we have observed that Traffic is marking both of the endpoints as degraded resulting in the traffic being routed to the passive node (which will be off and hence users will get error).

Pasted you below a snip of how this whole mechanism is setup

The public URL of traffic manager is mapped to two Public IPs which are associated to LBs running NVAs, then we have natted the traffic from endpoint to an internal app gateway which finally servers the app hosting VMs.

Now, for past few days we are observing that traffic manager is marking both of the endpoints as Degraded, even though backed infra and app is running fine, and app is responding back with 200 status code.

Any suggestions to direct us in a direction to investigate would be highly appreciated!

Thanks!

Answer 1

@Apurva Pathak ,

Welcome to the Microsoft Q&A Platform. Thank you for reaching out & I hope you are doing well.

I understand that you have two endPoints behind a Traffic Manager in an Active-Passive set up, but for some reason even the Active one is considered "Degraded" by Traffic Manager resulting in both the endPoints receiving traffic instead of Active only.

As next steps,

• Can you please check the Endpoint status by endpoint metric
  • Use "Figure 4" and apply filters to display individual metric
  • Do you see the active one being "0" or "1" ?
  • Please share a screenshot
• Can you confirm if the "Expected status code ranges" includes 200 ? Refer : Configure endpoint monitoring
• Make sure "Probing interval" is set to 30 to avoid Fast probing and effectively lowering the load on the EndPoint
• Additionally, you can consider Priority traffic-routing method to prioritize one backend over the other

Cheers,

Kapil