# Private metric discrepancy record

Original blank worksheet. A saved record is not a sent support request or a confirmed incident. Keep real sources private.

- Question and post/account local IDs:
- Source A: metric label / scope / filter / observed UTC time / post age:
- Source A state: number / blank / loading / error / unavailable:
- Source A value (only for an observed number):
- Source B: label / scope / filter / observed UTC time / post age:
- Source B state and observed value:
- Unit comparison: event / display / distinct account / other:
- Same post, author, scope and observation time established?
- Original evidence private location (omit personal data from a shared copy):
- One repeat: environment/version, unchanged conditions, actual result:
- Known limits / unresolved question / plausible alternatives:
- Minimal redacted support text, if needed:
- Actual submission status and receipt, only after a real send:

Do not replace unknown states with zero. Do not divide mixed-time values or claim an incident from a ratio alone. This template performs no upload and sends nothing.