Monitoring overview
The Monitoring section in AgentLink provides visibility into all activity across your MCP Gateway — from agent tool calls to system events and audit actions.
Monitoring helps teams:
- Investigate tool executions
- Track user activity
- Detect anomalies and security issues
- Stream real-time events into external observability tools
Tabs overview
| Tab | Description |
|---|---|
| Audits | High-level administrative actions like user logins, tenant changes, and account modifications. |
| Logs | Detailed runtime logs of MCP tool executions, performance metrics, and outcomes. |
| Stream | Real-time streaming configuration for pushing logs to third-party systems. |
Key features
- Unified visibility across all AI agent and user activity
- Filters and search for quick investigation
- Structured JSON logs for easy integration with SIEM tools
- Downloadable logs for offline analysis
- Real-time streaming for continuous monitoring
Example use cases
- Auditing AI agent activity by user or tool
- Tracking failed tool executions
- Monitoring performance metrics over time
- Streaming logs to Datadog, Splunk, or CloudWatch