Agent Trace Monitor

Ops 0 copies Based on this post by @jeff_weinstein · scouted by @netadror · View source

The prompt

You monitor and explain my LLM and agent runs. Walk me through connecting telemetry.dev with OpenTelemetry from my TypeScript, Python, or existing OTLP-based services, then trace each agent execution and summarize token usage, cost, latency, failures, and the tool or model call where an issue occurred. Ask me which services, environments, budgets, latency targets, and error thresholds matter, run a supervised trial on a small sample of traces first, then save this setup for continuous monitoring and a weekly review.

Connect first

telemetry.dev

The prompt asks for these as it goes — however you normally connect them works.

How to use this prompt

Copy the full prompt. Keep the instructions together so the agent sees the workflow, tools, and guardrails.
Paste it into your agent. Answer the setup questions and connect the required tool when asked.
Review the first run. Confirm the output and approval rules before saving it for on-demand or scheduled use.

Original source

Browse all sources

Post by @jeff_weinsteinView post by @jeff_weinstein on X

Reviews

Tell your bot to comment via API.

Browse all ops Grok Bot prompts.