Cloud Dashboard

The Hangar Cloud dashboard provides fleet-wide visibility into your MCP infrastructure.

Fleet Overview

The main dashboard shows all connected agents, their MCP servers, and current health status.

Agent List

Each connected agent reports:

• Status — Online, degraded, or offline
• MCP Servers — Count and state distribution (ready, degraded, dead)
• Tools — Total tool count across all MCP servers
• Uptime — Time since last restart
• Version — Agent version and configuration hash

MCP Server Health

Drill into any agent to see individual MCP server health:

• State machine visualization (COLD → READY → DEGRADED → DEAD)
• Circuit breaker status and failure counts
• Latency percentiles (p50, p95, p99)
• Tool call throughput

Audit Trail

The audit trail records every tool invocation across your fleet with identity tracking.

Filtering

Filter audit records by:

• Time range — Last hour, 24h, 7d, 30d, or custom
• Agent — Specific agent or all agents
• MCP Server — Specific MCP server
• Caller — Identity of the invoker
• Severity — Info, warning, error
• Event type — Tool call, lifecycle, policy violation

CEF Export

Export audit records in Common Event Format for compliance:

GET /api/observability/audit/export?format=cef&from=2026-03-01&to=2026-03-28

Supported formats: cef, json, csv.

Metrics Charts

Live metrics visualizations for:

• Tool call rate (calls/second)
• Error rate and circuit breaker trips
• Latency distribution
• MCP Server state distribution over time
• Top callers by volume

Policy Management

Available on Pro and Enterprise plans.

Create and push policies to agents in real time:

• Tool access filtering (allow/deny lists)
• Rate limiting per caller or MCP server
• Execution timeouts
• Required authentication levels

See Teams & Roles for permission management.