debug-traces

Diagnose OpenTelemetry trace ingestion failures and correlation errors in Coval workflows.

2|Updated Feb 17, 2026
One-click install
npx skills add https://github.com/coval-ai/coval-external-skills --skill debug-traces
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: debug-traces
Source: https://github.com/coval-ai/coval-external-skills/tree/main/skills/traces/debug-traces
Command: npx skills add https://github.com/coval-ai/coval-external-skills --skill debug-traces

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and scripts (resource) components.

What problem does it solve?

This skill resolves issues where OpenTelemetry traces are missing, sparse, duplicated, or incorrectly correlated within the Coval evaluation platform, ensuring reliable observability for AI agents.

Core Features & Use Cases

  • Boundary Identification: Systematically isolates failures between agent export, network transport, and Coval ingestion.
  • Connectivity Validation: Provides automated checks to verify API authentication and target ID validity.
  • Troubleshooting Matrix: Maps specific symptoms like missing spans or incorrect correlation headers to actionable fixes for Vapi, PSTN, and WebSocket integrations.

Quick Start

Run the debug-traces skill to identify why traces are missing for your current simulation by providing the simulation output ID and your API key.

Frequently Asked Questions about debug-traces

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

FAQPage Schema
How do I fix missing OpenTelemetry traces in my Coval evaluation workflows?

To fix missing OpenTelemetry traces in Coval evaluation workflows, you need to isolate failures between agent export, network transport, and Coval ingestion by validating trace boundaries and API authentication. Providing the simulation output ID and API key allows automated checks to identify the specific ingestion failure.

What causes incorrectly correlated spans in Coval agent-based simulations?

Incorrectly correlated spans in Coval agent-based simulations are typically caused by invalid header configurations and mismatched target IDs. Validating trace integrity against simulation output IDs and verifying conversation monitoring requirements ensures proper correlation for production-grade observability.

How do I troubleshoot OpenTelemetry trace ingestion failures for Vapi and WebSocket integrations?

To troubleshoot OpenTelemetry trace ingestion failures for Vapi and WebSocket integrations, map specific symptoms like sparse or duplicated traces to actionable fixes using a troubleshooting matrix to verify connectivity and header configurations. This systematically resolves issues across different integration protocols.

Does Coval trace debugging support PSTN network monitoring?

Coval trace debugging supports PSTN network monitoring by analyzing trace boundaries and network connectivity specifically for PSTN integrations. The troubleshooting matrix maps PSTN-specific symptoms to actionable fixes, ensuring reliable observability for those agent simulations.

Why are my OpenTelemetry traces sparse or duplicated during Coval simulation monitoring?

Sparse or duplicated OpenTelemetry traces during Coval simulation monitoring occur due to network connectivity issues and incorrect header configurations between agent export and Coval ingestion. Boundary identification helps systematically isolate whether the failure originates in transport or ingestion processing.