Agent Know-How

Databricks announced that MLflow is now a supported destination for native OpenTelemetry export from the Claude Platform. Traces of model and tool calls land beside existing MLflow data, so teams can find recurring failures, such as an agent rereading the same source, and test whether a change helps.

For builders: OpenTelemetry export means trace capture needs no custom SDK, and the backend stays swappable. This is a vendor announcement; check span coverage on your own agents.

Cloudflare's agent tracing adds invoke_agent → chat/execute_tool → tool_approval spans to existing Workers traces, keyed by agent name, agent ID, and conversation ID. Sessions replay turn by turn.

The caveats are the news:

• Payload recording defaults differ between its two supported SDKs (off in one, on in the other).
• The docs warn traces are not lossless; payloads may be truncated.
• From October 1, 2026, every span counts as a billable event.

Check defaults per integration: one stack can over-retain personal data while another drops the tool arguments a debugging session needs.

LangSmith traces voice agents built on Pipecat, LiveKit, OpenAI Realtime, and Gemini Live, capturing audio, STT and TTS latency, interruptions, and tool calls in one trace.

The same span discipline now covers a real-time, turn-taking interface. For voice agents, per-stage latency and interruption events are the debugging data a text trace never had.

Pond stores agent sessions from several machines in one S3 bucket with no database service to run, adds search over the archive, and exposes the archive to the agent as an MCP server.

It is the self-owned counterweight to vendor trace stores: the portable-file position plus the multi-machine collection and search it was missing. It treats sessions as an asset to keep, not a debugging byproduct that ages out with a retention window. It is an early Show HN project.

Langfuse v4 rebuilds both agent traces and evaluation results onto one immutable ClickHouse table, collapsing what were separate storage paths into one queryable store.

A score can now be joined back to the run that produced it without crossing systems. It is the same infrastructure hardening as LangSmith's SmithDB, aimed at unifying capture and evaluation rather than indexing traces alone.

LangChain's SmithDB builds a custom inverted index over object storage so agent traces can be full-text searched and JSON-filtered directly. It reports a 400 ms median (P50) query latency, even though each trace is a large, deeply nested JSON document.

At fleet scale, "traces are stored somewhere" is not enough: search over payloads is what makes them usable in an incident. This is the bar a self-hosted trace store has to meet.

LangSmith's Trajectories view shows a full agent session (user turns, tool calls, sub-agent handoffs) as one conversational thread instead of a tree of nested spans.

It trades the completeness of the span tree for a fast top-to-bottom read. It targets long-running sessions, where expanding every span by hand is the bottleneck.

Amazon OpenSearch Service's MCP Apps return interactive visualizations alongside an agent's text response, from one locally run MCP server. An alert investigation (trace lookup, log-pattern clustering, distributed-trace analysis, RED metrics, service-dependency mapping) happens in one IDE chat thread.

The visualization renders the actual OpenSearch query result, not a model-generated chart. An engineer can verify each step the agent claims without re-running the query in a separate dashboard.

LangSmith now positions itself as one debug console for coding agents from different vendors: Claude Code, Codex, Cursor, and Copilot. It shows tool calls, sub-agent handoffs, errors, cost, and retries in one place.

For teams whose engineers use several coding agents, the trace format matters more than which agent produced the trace.