Goal
Avoid future corruption from partial restarts.

Steps
1) Define and document:

• Which services may be restarted alone
• Which require a full stack restart
  2) Train support staff:
• “If KDS state looks wrong after a partial restart, do a full service restart.”
  3) Prefer:
• Full service restart during outages
• Avoid piecemeal restarts during active service
  4) Document:
• Services restarted
• Order of restart
• Resulting behavior

Artifacts to Capture

• Restart sequence used
• Services involved
• Screenshots of incorrect vs corrected state