Coding

Latitude Observability

Try it

Export OpenClaw OTLP traces to Latitude for open-source LLM observability and evaluation. Use when configuring diagnostics.otel to send OpenClaw model-call,...

What it does

Export OpenClaw OTLP traces to Latitude for open-source LLM observability and evaluation. Use when configuring diagnostics.otel to send OpenClaw model-call, run, tool, and message spans to Latitude's ingest endpoint.

The skill document

Latitude observability for OpenClaw

Latitude is an open-source LLM observability and evaluation platform. It ingests OpenTelemetry traces, so OpenClaw activity becomes spans you can inspect, search, and evaluate, with token usage, cost, and latency aggregated at every level of the trace.

OpenClaw already exports OTLP/HTTP protobuf through its built-in diagnostics-otel plugin. This skill configures that exporter to send traces to Latitude. There is no extra code to install.

When to use

Use this when you want OpenClaw model-call, run, tool, and message spans to land in Latitude, or when someone asks how to wire OpenClaw telemetry into Latitude.

Prerequisites

  • A Latitude account and API key. Sign up at console.latitude.so, or self-host.

  • A Latitude project slug.

  • Set the credentials in your environment:

    export LATITUDE_API_KEY=
    export LATITUDE_PROJECT=
    

Configure

Enable the diagnostics-otel plugin and point diagnostics.otel at Latitude's ingestion endpoint, authenticating with two headers. Add this to your OpenClaw config (~/.openclaw/openclaw.json):

{
  plugins: {
    allow: ["diagnostics-otel"],
    entries: {
      "diagnostics-otel": { enabled: true },
    },
  },
  diagnostics: {
    enabled: true,
    otel: {
      enabled: true,
      endpoint: "https://ingest.latitude.so",
      protocol: "http/protobuf",
      serviceName: "openclaw-gateway",
      headers: {
        Authorization: "Bearer ${LATITUDE_API_KEY}",
        "X-Latitude-Project": "${LATITUDE_PROJECT}",
      },
      traces: true,
      metrics: false,
      logs: false,
    },
  },
}

For packaged installs, install the plugin first:

openclaw plugins install clawhub:@openclaw/diagnostics-otel

Notes

  • OpenClaw appends /v1/traces to the base endpoint, so set endpoint to https://ingest.latitude.so (no path). If you self-host Latitude, point it at your own ingestion host instead.
  • protocol must be http/protobuf. OpenClaw ignores grpc today, and Latitude ingests OTLP/HTTP protobuf.
  • Latitude ingests OTLP traces, so this config leaves metrics and logs off. Route those signals to a separate metrics or logs backend if you need them.
  • Raw prompt and response content is not exported by default. Enable diagnostics.otel.captureContent.* only when your retention policy allows it.

Verify

Run an OpenClaw agent turn that calls a model and a tool, then open your project in the Latitude dashboard. You should see a trace with nested spans for the run, model call, and tool call, with token usage, cost, and latency aggregated per trace.

Related skills

Connect OpenClaw to the Latitude MCP server so the agent can read and manage your Latitude workspace (projects, traces, spans, issues, annotations, scores, s...

1 installs

Smart LLM Router for OpenClaw. Save up to 70% by routing every request to the right model. No coding required.

13 installs

AI 应用可观测性与生产监控实操手册——AI 进生产后的"仪表盘与探照灯":可观测性全景(三大支柱:日志/指标/追踪 + AI 特有观测对象)、调用追踪与日志规范(LLM 调用追踪、Span 设计、会话级追踪、敏感信息过滤)、质量监控指标(幻觉率/拒答率/满意度/转人工率实时看板)、性能与成本监控(延迟/吞吐/Token 成本实时追踪)、护栏与安全监控(护栏命中率/注入检测/敏感数据泄漏监控)、告警体系(分级告警/阈值设计/通知路由/告警疲劳治理)、监控平台与落地(埋点规范/工具选型/灰度期监控)。附零依赖本地工具一键出三大支柱清单、监控指标表、告警设计、追踪规范与落地路线。面向 AI 平台、SRE、运维与质量负责人——与 LLM 评测(离线质量)互补,本技能管线上运行质量。

Specify the tracing, metrics, and alerting for an AI agent or LLM feature in production. Use when asked what to log for an LLM app, design agent tracing or s...

Track LLM API spending per agent/session with budget alerts and CSV export

1 installs

Record every tool call into agent-tracer (self-hosted FastAPI), then run regression testing (golden case → drift detection), attribute token costs by tool/model/agent, and surface recurring error root causes across sessions. Use when an agent must log its actions, verify behavior hasn't drifted, report spend, or debug repeated failures.