The problem
Backup checks lived in emails and memory. There was no single place to see whether last night’s backups worked, show evidence that checks were done, or track a failure until it was fixed. My own work history was scattered too.
What I did
- Designed and built a single-container web app with a status page filled automatically each day by an evidence-collection script, showing job results, offsite-drive rotation, storage capacity and offsite copies. Data older than a day is flagged as stale, so a broken collector can’t look like a clean night.
- Built weekly, monthly and quarterly verification checklists: each run records pass, fail or not applicable per item, with notes and evidence uploads, and calculates the overall result. Late checks are flagged as overdue.
- Added a findings register that tracks each failure to closure, with a severity, an owner and a due date.
- Added a backup events log, container-host status, and period reports with CSV and Markdown export for the audit trail.
- Added a daily work log with an ingest API, so completed work is recorded as it happens, by me or by Claude.
- Hosted it away from the backup server on purpose, so the evidence survives if the backup server doesn’t.
Outcome
One page answers “did backups work?”, every check and finding leaves evidence, and there’s a searchable record of every day’s work.