Insights & Observability
Insights is the portal's observability hub. It gives you a unified view of all telemetry data from your applications — logs, traces, metrics, events, messages, and audit events.
Panels
Insights is organized into panels, each focused on a specific data type:
| Panel | Description |
|---|---|
| Overview | Stat cards, sparkline charts, and recent activity |
| Conversations | Messages exchanged between agents and services |
| Events | Application events and workflow triggers |
| Logs | Log entries with severity filtering and search |
| Traces | Distributed trace spans with waterfall visualization |
| Metrics | Time-series metric charts and raw data |
| Audit Trail | Security and compliance audit events |
| Network | Active NATS topics scoped to your organization |
The Data Pipeline
Your telemetry data flows through a pipeline before appearing in Insights:
Your App → Air API (ingest) → NATS → Pulse (Parquet writer) → Insights (SQL query)- Your app sends data via the Air API (HTTP)
- Data flows through NATS for real-time delivery
- Pulse writes data to Parquet files (durable storage)
- Insights queries Parquet files using DataFusion SQL
Common Filters
Every panel supports a time range selector:
- Last hour — Most recent data
- Last 24 hours — Default view
- Last 7 days — Weekly overview
- Last 30 days — Monthly overview
Each panel also has type-specific filters (severity, status, search, etc.).
Cross-Panel Navigation
Insights panels are linked together for correlation:
- In Logs, click a trace ID to jump to the Traces panel filtered to that trace
- In Traces, click "View Logs" on a span to jump to Logs filtered by trace ID
- In the Overview, click a stat card to jump to the corresponding panel
- URLs include query parameters (e.g.,
?trace_id=abc123) so filtered views are bookmarkable
Guides
- Overview Dashboard — Stat cards and sparklines
- Logs — Search, filter, and correlate logs
- Traces — Waterfall visualization and span details
- Metrics — Time-series charts
- Audit Trail — Security and compliance events
Next Steps
- Send telemetry data from your apps
- Query with SQL for advanced analysis
- Build a dashboard with custom charts