Langfuse v4.45.2
Langfuse v4.45.2 introduces project-level batching and transient failure retries for AI Gateway inference telemetry, along with OpenAPI typing fixes and web UI chart refactors.
原始内容为英文；当前页面提供中文导航与来源说明，具体事实请以原文为准。
为什么值得关注
Langfuse v4.45.2 delivers a set of features, fixes, and refactors across backend and frontend components. Key updates include: - AI Gateway: Implements batch inference telemetry per project and adds retry mechanisms for transient failures ( 17855). - Bug Fixes: Retains metadata keys conflicting with Object.prototype ( 17686), replaces generic S3 errors with underlying read failures in batch export ( 17891), properly types OtelSpan scalars and dataset JSON-schema fields in OpenAPI ( 17892), resolves session metrics bar z-index issues ( 17887), and omits IdP-reported issuers from SSO discovery error outputs ( 17880). - Performance & Refactoring: Dedups score events rows path using GROUP BY + INNER JOIN ( 17883), raises index project trace observation granularity ( 17882), unifies cursor pagination metadata into a shared API envelope ( 17888), and replaces the recharts dependency with design-system charts in score analytics ( 17885, 17889).
可执行摘要
Langfuse released v4.45.2 featuring per-project telemetry batching and retry logic for transient errors in the AI Gateway, alongside fixes for Object.prototype metadata key collisions, accurate batch-export S3 error reporting, OpenAPI type definitions for OtelSpan and dataset schemas, and database query optimizations for scores.
- Agent 实用度
- 82/100
- 可信度
- 96%
- 机器格式
- JSON + Markdown
开发者应核对什么
- Upgrade Langfuse server instances to v4.45.2.
- Regenerate client SDKs if relying on OpenAPI types for OtelSpan scalars or dataset JSON-schema fields.
- Verify AI Gateway configuration to ensure project-level batch telemetry and retry settings align with throughput requirements.