heady-observability-mesh

Unify Langfuse, Sentry, and OpenTelemetry tracing across multi-provider LLM pipelines.

1|Updated Mar 24, 2026
One-click install
npx skills add https://github.com/HeadyAI/heady-context --skill heady-observability-mesh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: heady-observability-mesh
Source: https://github.com/HeadyAI/heady-context/tree/main/heady-skills/heady-observability-mesh
Command: npx skills add https://github.com/HeadyAI/heady-context --skill heady-observability-mesh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Unified AI observability across Langfuse tracing, Sentry error tracking, and OpenTelemetry distributed tracing for the Heady ecosystem.

Core Features & Use Cases

  • Trace correlation: single correlation_id propagated from the edge through Cloud Run, Neon, Upstash, and LLM providers to unify end-to-end traces.
  • OpenTelemetry per-stage spans: 21 HCFP stages wrapped with phi-aware timing and tags for governance-level observability.
  • Langfuse tracing & cost analytics: per-LLM-call tracing, token accounting, and provider-cost aggregation.
  • Sentry error tracking with CSL: automatic classification and tagging by Sacred Geometry context for proactive alerting.
  • Coherence scoring & alerts: phi-weighted coherence across latency, errors, and cost with threshold-based alerts.
  • Health and API surface: /health, /trace/llm, /coherence, and /cost endpoints for operational visibility.

Quick Start

Integrate the heady-observability-mesh router into your service and verify readiness with the /health endpoint.

Frequently Asked Questions about heady-observability-mesh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I unify OpenTelemetry distributed tracing with Sentry error tracking and LLM tracing?

Unify OpenTelemetry distributed tracing, Sentry, and LLM tracing by propagating a single correlation_id across Cloudflare edge, Cloud Run, and Neon, linking per-stage spans with LLM cost analytics and classified errors.

How does correlation context propagation work for multi-provider LLM pipelines?

Correlation context propagation works by passing a single correlation_id from the edge through Cloud Run, Neon, Upstash, and LLM providers, ensuring end-to-end trace continuity across multi-provider pipelines.

Can I track per-LLM-call cost analytics and token usage with OpenTelemetry?

Yes, you can track per-LLM-call cost analytics by wrapping LLM interactions with Langfuse tracing, providing token accounting and provider-cost aggregation alongside OpenTelemetry per-stage spans.

Does this observability mesh support alerting for latency and coherence thresholds?

The observability mesh supports alerting by calculating phi-weighted coherence scores across latency, errors, and cost, triggering threshold-based alerts for operational visibility.

What endpoints are available for operational visibility in an AI observability mesh?

Available endpoints for operational visibility include /health for readiness checks, /trace/llm for LLM traces, /coherence for scoring, and /cost for provider-cost aggregation.

How do I classify Sentry errors using Sacred Geometry context for proactive alerting?

Classify Sentry errors automatically by applying CSL-based error tagging with Sacred Geometry context, enabling proactive alerting and structured error tracking across distributed traces.