Source: github.com/langgenius/dify/releases/tag/1.19.0
Dify 1.19.0, published by the langgenius team on 4 November, extends tracing to cover Workflow Tools that an outer workflow calls. External traces now include the persisted execution steps inside those tools, nested under the Tool span that invoked them. The release lists Langfuse, LangSmith, Opik, Weave, Arize, Phoenix, Aliyun, MLflow, Databricks and Tencent among the providers that share this export path.
Exported app packages can now include the workflows behind their Workflow Tools, including transitive dependencies. On import, matching tools already in the destination workspace are reused and missing ones are recreated and published. All app exports now use package format v2, and version 1 imports remain supported. The release also adds difyctl commands for listing, exporting, importing and publishing workspace Skills.
The upgrade requires Docker Compose 2.35.0 or later, and the bundled plugin daemon moves to 0.6.14-local. Existing tracing history is not rebuilt, so older exported traces will not gain the new nested spans. Rolling deployments need API, workers and Beat to agree on the new ops_trace_v2 tracing queue.
Why it matters if you build agents
Developers who debug agents in Langfuse or similar tools can now follow a run into a nested Workflow Tool call, including repeated and failed invocations. Teams that move apps between workspaces should confirm the destination accepts format v2 packages before importing, since older importers may reject them. Rolling upgrades should keep the old tracing workers running until queued work on the old queue has drained.
Get stories like this every weekday
One short email with the day's agent tooling news.