GOVERNATLAS / PRODUCT DOSSIER

AgentOps

AgentOps v2 documentation describes SDK-based recording of traces and sessions, LLM calls, tool calls and errors, with session drilldown and waterfall views. It also describes manual trace control; complete automatic instrumentation and production reliability are not independently established.

Explore the product ↗Add to comparison ↗

01 / CONTEXT & SCOPE

Understand the proposition.
Define the boundary.

The decision to make

Assess whether the session view explains the failure your team actually needs to debug. Define an expected event sequence before instrumentation: user turn, model call, tool call, retry and final response. A long waterfall may still omit the custom operation that caused the incident, so compare it with an application-owned event log.

What this record does not establish

The captured v2 documentation supports a scoped description of SDK recording, sessions, drilldown and manual controls. It does not prove that every framework event is captured or that raw prompt and completion data are appropriate to retain. Confirm SDK compatibility, sampling, redaction, access and retention for your deployment before sending sensitive traces.

1 current source · 1 published assertion
Source coverage is a research boundary, not a quality score.

02 / YOUR EVALUATION

Turn interest into a useful trial.

Open the evaluation playbook ↗

A test worth running

Run a proposed sandbox conversation with a successful tool call, a tool exception, a retry and a custom operation outside automatic instrumentation. Compare session and trace boundaries, timestamps and identifiers with the application's record. Add manual tracing for the missing operation and repeat. Ask an operator to explain what changed without relying on an unrecorded terminal session.

GovernAtlas editorial proposal. We have not performed this test.

Evidence to bring back

An application-log versus session comparison identifying missing, duplicated and manually instrumented events.
A reviewed trace-data inventory and operating record for SDK versions, access permissions, retention and failed ingestion.

03 / DOCUMENTED EVIDENCE

Read the claim. Inspect the basis.

Each statement retains its classification and preserved source. A vendor assertion is not an independent test result.

vendor asserted

Source, dates & preserved provenance

AgentOps — official project or product documentation ↗

Authority: vendor owned
Observed 2026-10-15 · Reviewed 2026-10-15 · current until 2026-11-14

Artifact SHA-256: b44db1669ed14fa9192832748ba478328d7ccca10451bf5a739611f015880aea

The hash identifies the preserved capture. The live source may have changed.

Source register

AgentOps — official project or product documentation ↗vendor owned · captured 2026-10-15

Brand asset provenance

Organization identity asset from its official website; inclusion does not imply endorsement.

Official website ↗ · Original asset ↗

04 / BEFORE YOU COMMIT

Better questions.
A clearer decision.

Resolve these with the provider and your own evaluation team.

Which agent, tool and error events remain missing after instrumentation, and how are they reconciled with application logs?

Which deployment-specific evidence establishes access, retention, export and deletion behavior?

What happens when capture or evaluation fails, and who owns the unresolved gap?

Regulatory applicability requires a separate, current determination. Inclusion is not certification.

KEEP THE RECORD USEFUL

A dossier should stand up to scrutiny.

Have a source that changes this picture? Submit it for editorial review.

Raise an evidence issue ↗