MLOps Academy · 课时

跟踪并监控 LLM 调用

记录每个请求的令牌数、延迟和成本

跟踪并监控 LLM 调用 是 CoddyKit 上的免费 MLOps Academy 课时。这是第 3 节课，共 4 节。你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。这是 MLOps Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 MLOps Academy 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

You Cannot Fix What You Cannot See

An LLM app feels like a black box until you log what it does. Tracing records each call so you can debug, cost, and improve it later. 🔍

Capture the Request

For every call, log the final prompt, the model name, and key parameters. This request record lets you reproduce any answer exactly.

log.info({"model": "gpt-4o", "prompt": prompt, "temperature": 0.2})

Capture the Response

Save the generated text alongside the request. Together the input and output form one trace you can replay, score, or share with a teammate.

Record Token Usage

Each response reports tokens in and out. Logging tokens per call is how you attribute cost and spot prompts that quietly grew too large.

usage = response.usage
log.info({"in": usage.input_tokens, "out": usage.output_tokens})

Turn Tokens Into Cost

Multiply token counts by the price per token to get spend. Tracking cost per request keeps a runaway feature from surprising you on the bill.

cost = inp / 1e6 * 2.50 + out / 1e6 * 10.00

Measure Latency

Time each call from send to final token. Watching latency per request reveals slow prompts and tells you when streaming would help users.

start = time.perf_counter()
response = client.responses.create(...)
latency = time.perf_counter() - start

Spans Build a Trace

A real request may chain retrieval, the LLM, and a tool. Each step is a span, and nesting them into one trace shows where time and tokens go.

Tools Built for LLMs

Platforms like LangSmith, Langfuse, and Phoenix capture traces with one wrapper. A purpose-built observability tool beats parsing raw logs by hand.

from langfuse.openai import openai
response = openai.responses.create(model="gpt-4o", input=prompt)

Tag Traces With Metadata

Attach a user id, prompt version, and feature name to each trace. This metadata lets you slice metrics and find which segment is misbehaving.

Sample What You Cannot Store

At high volume, logging every full payload is costly. Sampling a fraction of traces keeps insight while controlling storage and privacy.

Dashboards and Alerts

Roll traces into dashboards for cost, latency, and error rate, then alert on spikes. Now production problems reach you before users complain.

Quick Check

Your monthly LLM bill jumped with no traffic change. Which logged value explains it fastest?

Recap

You can now trace LLM calls end to end: request, response, tokens, cost, latency, and spans, then watch them on dashboards. Eyes wide open! 🎉

常见问题解答

「跟踪并监控 LLM 调用」课时是免费的吗？

是的 — 「跟踪并监控 LLM 调用」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 MLOps Academy 课程的其余内容，请升级到 CoddyKit PRO。 MLOps Academy 课程共包含 4 节课。

「跟踪并监控 LLM 调用」这节课中我会学到什么？

记录每个请求的令牌数、延迟和成本你通过在浏览器中直接运行的动手代码来练习 MLOps Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 MLOps Academy 需要有经验吗？

无需任何先前经验。CoddyKit 上的 MLOps Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。这是第 3 节课，共 4 节。

「跟踪并监控 LLM 调用」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 MLOps Academy 课中编写并运行代码吗？

能。每节 MLOps Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

← 返回 MLOps Academy