Artificial Intelligence1 min read

Claude Platform now supports native OpenTelemetry observability in Databricks MLflow

By · Published by Everything Blog

In short

The integration relies on Databricks Zerobus Ingest to capture OpenTelemetry spans at high query rates with low latency. In addition to monitoring execution timing, model requests, and tool calls, organizations can activate a "full" content mode to record complete inputs and outputs. Development teams can then leverage MLflow's Analyze trace action alongside the beta MLflow Issues tool, which both use Genie Code to troubleshoot operational failures and recurring errors.

Key points

  • Genie Code powers the MLflow Analyze trace capability to automatically explain complex ex…: Genie Code powers the MLflow Analyze trace capability to automatically explain complex execution errors.
  • The beta MLflow Issues feature allows engineers to detect recurring operational defects l…: The beta MLflow Issues feature allows engineers to detect recurring operational defects like repeated searches or broken tool calls.
  • Storing traces in Unity Catalog allows enterprises to execute SQL queries and enforce str…: Storing traces in Unity Catalog allows enterprises to execute SQL queries and enforce strict data access governance.

Sources checked by our analysis pipeline: 1

Unify Claude observability into Databricks with native OpenTelemetry ingestion. Use collected traces with MLflow to find recurring failures and improve the agents.

An agent can return a plausible answer while missing an important finding or reading the same source repeatedly. The answer alone won't explain why. Traces show the model and tool calls behind it, so you can investigate the failure, find related cases, and test whether a change helps.

Databricks MLflow is now a supported destination for native OpenTelemetry export from the Claude Platform.. With native export to Databricks MLflow, developers using the Anthropic SDKs, the Messages API and Managed Agents can send native traces to Databricks and use MLflow to debug their agents and evaluate changes.

MLflow Tracing is built on OpenTelemetry, an open standard for collecting and exporting service signals. A trace records an agent’s execution as individual steps, called spans, such as model requests and tool calls. Claude’s native instrumentation generates these spans and exports them using the OpenTelemetry Protocol.

Databricks receives the spans through Zerobus Ingest and stores them in Unity Catalog tables linked to…

Artificial Intelligence

More in Artificial Intelligence