Configuring trace connections

  • Release version: Zurich
  • Updated June 30, 2026
  • 1 minute to read

    • Collect trace data from AI agents running on supported cloud platforms and monitoring services to generate evaluation metrics, populate AI inventory, and support security monitoring in AI Control Tower.

    Key benefits

    • Collect trace data from AWS, Google Cloud, and Microsoft Azure platforms through a single interface, without SDK instrumentation in your agent code.
    • Discover AI agents running on third-party platforms and add them to your AI asset inventory automatically.
    • Generate evaluation and security metrics from collected traces when evaluations and security monitoring capabilities are enabled.
    • Access cloud platforms securely through a MID Server, which authenticates to your cloud provider and keeps credentials protected within your ServiceNow instance.

    Trace connections and SDK instrumentation

    AI Control Tower supports two ways to collect trace data from AI systems: a trace connection, which polls a supported cloud platform at a configured interval, or SDK instrumentation, which sends trace data using an API key. Use a trace connection when your AI systems run on a supported platform.

    For AI systems hosted on other platforms or third-party agents, send trace data using the API instead. See Activate evaluation scoring for external AI systems.

    Trace connection limitations

    • Up to 5 trace connections for each connector run in parallel. You can create more than 5, but only 5 run at a time.
    • Manually running a trace connection, instead of waiting for its scheduled interval, counts against a separate limit: up to 2 manual runs at a time.