Description of problem: It was reported earlier today that routes were taking at least an hour to become functional on starter-us-east-1. During my testing, 20 out of 20 test apps still are reporting 503s when I try to curl the route, even 30+ minutes after the app started Running. It's unknown how long routes will take to start working (if ever), but there is a very long delay between when a pod enters Running and state when the route will return a 200. Version-Release number of selected component (if applicable): v3.5.5.19 How reproducible: Every time, on this cluster https://console.starter-us-east-1.openshift.com Steps to Reproduce: 1. Create a project. 2. Wait 5 minutes before creating an app, because of this missing token bug https://bugzilla.redhat.com/show_bug.cgi?id=1462542 3. Create an app in the new project. Actual results: The route does not return a 200. It returns 503 for a very long time. Expected results: The route should return 200 as soon as the backing pod is ready and Running. Additional info:
If you are using https://github.com/openshift/origin/blob/master/examples/quickstarts/cakephp-mysql-persistent.json then you need to use http not https to hit the routes.
Or create/edit the route with edge termination oc patch route -n ${namespace} ${route} -p '{"spec":{"tls": {"termination": "edge", "insecureEdgeTerminationPolicy": "Allow"}}}'