Following a thread with rhn-gps-rspazzol and rhn-gps-scworthi regarding an unusual failure, we worked out quickly that the problem was due to the security setting in the Jaeger CR being reset to none. rcernich1 has reviewed the operator code and found a sequence of events which explains the behaviour when running within OpenShift.

There are two times when the platform is 'detected'

• on startup when the proper detection code is executed
• in response to an error during the 5 second periodic checks

It is the second of these which is problematic since any error returned by the discovery client will result in the platform being permanently reset to k8s which then leads to the security being reset to None.