Agent911 — Reliability Snapshot Explained

Agent911 is the control plane for agent reliability — a read-only, unified surface for system stability, protection history, and operator visibility. It aggregates signals, highlights risk state, and provides guided triage. Agent911 v0.1 does not perform recovery on its own.

What a "Snapshot" Means

A snapshot is a unified view of:

• current health signals (liveness, silence, stalls)
• recent anomalies and correlations (including compaction pressure)
• routing / governance status (when available)
• recovery readiness indicators (backup posture, restore assumptions)

It is designed to answer:

"What is happening right now, and what should I do next?"

What Agent911 Does

• Aggregates and surfaces reliability evidence
• Provides deterministic recovery guidance
• Produces operator-ready proof bundles (logs/excerpts, event traces)
• Maintains a system-of-record view for incidents

What Agent911 Does NOT Do (v0.1 Truth)

• No automatic recovery
• No automatic repairs
• No config mutation by default
• No "magic fixes" without operator intent

When to Use Agent911

Use Agent911 when:

• the gateway is "up" but behavior is stuck
• the system is slow, flaky, or unresponsive
• you need a unified incident view for decision-making
• you need a repeatable recovery playbook and proof artifacts

Next Step

If you're new, start with the 5-minute path:

• /docs/quickstart/5-minute