Collect status info where necessary regardless of config

• Collect status info where necessary regardless of config
• If response messages are persistent then
  • use a probably faster mechanism on startup to clear queues
  • give a warning