Skip to content

Langfuse 4.72.0 adds gateway analytics and PagerDuty automation

Updated Langfuse

Summary

Gateway attribution, trace-level evaluation filters and notification routing expand the observability workflow.

Category
Agent infra
Date
Pricing
Not stated in the sources we cited.

Langfuse 4.72.0 adds an analytics tab in AI Gateway settings and attributes gateway spans to their API key and service. It also updates onboarding and user attribution for generated keys.

Route events into operations

PagerDuty integration can send alerts, project notifications and prompt changes, with the routing key configured on the automation. This is a notification integration, not evidence that an alert reaches an on-call responder in every deployment.

Evaluation filters gain trace-level conditions. This lets teams target evaluation rules using trace context rather than treating every trace as equivalent.

Self-hosted behavior to review

The release adds reporting of AI Gateway request counts from self-hosted instances. Operators should review that telemetry change alongside their deployment settings. It also corrects double counting of cache-write TTL usage in generic OpenTelemetry mapping. Accounting corrections should not be mistaken for a change in underlying model-provider prices.

See the langfuse hub.

More from Langfuse

View all

Questions

What is Langfuse?

According to its site, Langfuse is an open source agent evals and observability platform to trace, evaluate and improve AI agents.

Is Langfuse open source?

The product site says all product features are MIT licensed, though the GitHub API reports no recognized license for the repository.

Does Langfuse support MCP?

The product site says SKILL.md, CLI and MCP connect coding agents to Langfuse.

Filed on Agent Radar. Claims in this note are limited to the sources above.