Participant uptime is shared operational truth

What the public implementation shows

Guevara stores MonitHistory, exposes institution uptime queries, and includes monitoring-history cleanup and participant-status workers.

External monitors provide observations while the control plane preserves participant identity, history, and governed access. This is the implementation fact behind the note; the architectural reading begins from that visible boundary rather than from an assumed production story.

Operating principleAvailability governance requires a historical record, not a screenshot of current state.

The decision behind the mechanism

Historical ingestion separates the monitoring tool from the institutional evidence used for SLA and incident review.

Current status alone cannot explain duration, recurrence, observation gaps, or whether a participant recovered within expectation. The mechanism matters because it gives that failure a named boundary, durable evidence, and a controlled response instead of leaving the outcome to timing or operator guesswork.

• Retain observation time and monitored component.
• Distinguish down from unknown or missing data.
• Aggregate only over a declared observation policy.
• Apply retention without erasing required incident evidence.

How the control should be operated

Uptime views should disclose observation interval, missing-data policy, component scope, retention, and the threshold used for classification.

Participant conversations can use shared evidence instead of competing screenshots. The reusable lesson is not to copy a class or endpoint in isolation, but to preserve the relationship between identity, state, authority, evidence, and operational ownership.