A reporting tool should help someone decide whether to trust an observation, request a new measurement or investigate a gap. That is the purpose of this small, dated demonstration.
The collection/archive and logical-restore entries summarize historical events reported in retained operational records. Their original private receipts are not included in this download, so the package demonstrates how those reported facts are handled, not independent proof that the events occurred. The capacity entry uses a dated example and an illustrative seven-day freshness window. The missing independent-device verification is deliberately unknown.
The broader system is Ivy Control, a private operational system. Ivy Health Dashboard is this distinct reporting product. It has no live connection, scheduled refresh, alerts or claim of current service health.
| State | Meaning |
|---|---|
| Current | A supplied observation is at or within its stated age limit. This describes freshness, not overall health. |
| Stale | The observation is older than its limit. Request a new measurement. |
| Historical | A dated event remains a past event; it cannot age into proof of current operation. |
| Unknown | No qualifying observation was supplied. Configuration or a different successful step is insufficient. |
The Python builder validates a deliberately small input format, rejects malformed or future observation dates and unsupported fields, and calculates age against the explicit assessment time. Fractional seconds count. Historical and unknown entries cannot carry freshness windows. The normalized JSON retains the applied window and age so readers can check the calculation.
Input is curated by a person. Structural checks and a limited text denylist are defense in depth, not automatic redaction or proof that a statement is true. The code cannot authenticate sources or decide whether an event was adequately verified. It escapes text before rendering it as HTML.
Download the standalone source, example and tests · Inspect the displayed evidence JSON · Read the reproduction guide
python3 tools/public_health_dashboard.py --input tests/fixtures/ivy_public_dashboard_evidence.json --output-dir out python3 -m pytest -q tests/test_public_health_dashboard.py
Python 3.11+ builds the six files without network access or dependencies. Tests require pytest. The package includes synthetic adversarial cases, not private operational receipts. Tests verify rules, boundary conditions, HTML escaping and generation; they do not verify service uptime or a real recovery.
A logical restore can establish that a specified database backup reconstructs the checked data and structure. It does not establish recurring backups, a second device, cluster-wide roles, production privileges, whole-machine recovery or current recovery readiness. A disk-space observation is likewise evidence about a moment, not a guarantee of continuing capacity.
Buddy designed this evidence-reporting surface using AI-assisted engineering and review. Product choices, validation and publication remain human responsibilities. The owned renderer, tests, documentation and curated example in this standalone package are licensed under the MIT License. Private receipts, third-party material and the broader operational system are outside this grant.