I'm trying to troubleshoot the health of our origins in Azure Front Door. The main issue is that we sometimes see that some requests are routed to the second priority even though session affinity is enabled. My thinking is that, for some reason, the primary origin is seen as unhealthy by Azure Front Door.
When I drill into the metrics I can see that we don't have a 100% health report across the services and that there are some occasional dips:
But when I dig into the probe logs I'm not seeing as many errors as I would have expected:
How do I go about troubleshooting this? It appears that the metrics could be in line with my theory as to why the secondary origin is being hit but I feel like some logs are missing. I would have liked to see a steady 100% graph instead of the dips.