Print for desired state could show too much information from flinkProperties

[a-roberts]

Hi everyone,

We currently have these log statements

*desired = getDesiredClusterState(observed.cluster, time.Now())
    if desired.ConfigMap != nil {
        log.Info("Desired state", "ConfigMap", *desired.ConfigMap)
    } else {
        log.Info("Desired state", "ConfigMap", "nil")
    }
    if desired.JmDeployment != nil {
        log.Info("Desired state", "JobManager deployment", *desired.JmDeployment)
    } else {
        log.Info("Desired state", "JobManager deployment", "nil")
    }
    if desired.JmService != nil {
        log.Info("Desired state", "JobManager service", *desired.JmService)
    } else {
        log.Info("Desired state", "JobManager service", "nil")
    }
    if desired.JmIngress != nil {
        log.Info("Desired state", "JobManager ingress", *desired.JmIngress)
    } else {
        log.Info("Desired state", "JobManager ingress", "nil")
    }
    if desired.TmDeployment != nil {
        log.Info("Desired state", "TaskManager deployment", *desired.TmDeployment)
    } else {
        log.Info("Desired state", "TaskManager deployment", "nil")
    }
    if desired.Job != nil {
        log.Info("Desired state", "Job", *desired.Job)
    } else {
        log.Info("Desired state", "Job", "nil")
    }

    log.Info("---------- 4. Take actions ----------")

and it's possible a user puts something sensitive (like trust store passwords) into the flinkProperties, for their FlinkCluster - and thus it would be exposed through the pod logs. I think these statements should be removed - thoughts? Yeah it's useful for debug but shouldn't really be required I think, best to be on the safe side!

[functicons]

Good point! I think we can add a command line flag or add a field in CRD, and turn off these logs by default.