Observability & traceability

Datadog Agent Observability

Datadog's AI observability product, documented as span kinds for LLM, workflow, agent, tool, task, embedding and retrieval. LLM spans record model and provider, session ID, inputs and outputs, tool calls, results and definitions, token and cost metrics, metadata and errors, inside an enterprise platform with audit trail, SAML SSO and access control.

commercial · generally available · Research snapshot 2026-09-06

Visit the official product source ↗

Where it fits

Observability & traceability

Useful conversation with: SRE lead, Platform engineering manager, CISO.

Ask for a demonstration

Show me an agent span tree with tool calls and token cost metrics, and how retention plus RBAC apply to the captured prompt and response data.

Capabilities and evidence

Support labels reflect the supplied research. Documentation and vendor claims are not independent product tests. “Not established” means the researcher did not find support; it does not prove a capability is absent.

Documented by provider

Setup documentation requires a span kind from LLM, workflow, agent, tool, task, embedding and retrieval, and states spans with an invalid span kind are not submitted.

Limit: Multi-agent handoff is not named as a first-class concept; it must be modelled with agent and workflow spans.

Source s2

Documented by provider

LLM spans capture invoked model name, operation name, model provider, user session ID, ML application name, input and output data, token metrics, cost metrics, metadata, tags, tool calls, tool results, tool definitions, prompt metadata and errors, with named token and cost metric fields; agent and tool spans capture function inputs, outputs, tags, metadata and errors.

Limit: Manually instrumented spans require the customer to record token counts, so completeness depends on instrumentation quality.

Source s2

Documented by provider

Product navigation includes a Data Privacy, Security and RBAC section, and platform administration documentation surfaces Audit Trail, Access Control, SSO with SAML and Data Security.

Limit: The pages fetched only expose navigation labels; retention periods, PII scrubbing behaviour and whether Audit Trail covers AI telemetry access are not established.

Source s1 · Source s3

Limitations to discuss

Sources

  1. Agent Observability · Datadog · official docs
    Access date reported by researcher: 2026-09-06
  2. Set up LLM Observability · Datadog · official docs
    Access date reported by researcher: 2026-09-06
  3. Agent Observability Instrumentation · Datadog · official docs
    Access date reported by researcher: 2026-09-06

Listing does not imply partnership, supplier status, a working DutyGraph integration, or a compliance certification.

Suggest a correction