How to beginner · 3 min read

How to log AI workflow execution

Q: How to log AI workflow execution

Use Python's logging module to capture AI workflow execution details such as inputs, outputs, timestamps, and errors. Integrate logging calls around your OpenAI SDK calls to record each step of the AI interaction for traceability and debugging.

Quick answer

Use Python's logging module to capture AI workflow execution details such as inputs, outputs, timestamps, and errors. Integrate logging calls around your OpenAI SDK calls to record each step of the AI interaction for traceability and debugging.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup logging in Python

Configure Python's built-in logging module to output logs to console or file with timestamps and log levels. This setup is essential for capturing AI workflow events.

python

import logging

logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    handlers=[logging.StreamHandler()]
)
logger = logging.getLogger(__name__)

logger.info("Logging is configured.")

output

2026-04-27 12:00:00,000 - INFO - Logging is configured.

Step by step AI workflow logging

Wrap your AI calls with logging statements to record inputs, outputs, and errors. This example uses the openai SDK v1+ pattern with OpenAI client and gpt-4o model.

python

import os
from openai import OpenAI
import logging

logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

try:
    user_input = "Explain AI workflow logging"
    logger.info(f"User input: {user_input}")

    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[{"role": "user", "content": user_input}]
    )

    output_text = response.choices[0].message.content
    logger.info(f"AI output: {output_text}")

except Exception as e:
    logger.error(f"Error during AI call: {e}")
    raise

output

2026-04-27 12:00:01,000 - INFO - User input: Explain AI workflow logging
2026-04-27 12:00:02,500 - INFO - AI output: AI workflow logging involves capturing inputs, outputs, timestamps, and errors to monitor and debug AI applications.

Common variations

You can extend logging to async workflows, streaming responses, or use other AI SDKs like Anthropic or Google Vertex AI. For async, use asyncio and async for with logging inside the loop. For streaming, log each chunk received.

python

import os
import asyncio
from openai import OpenAI
import logging

logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

async def async_stream_log():
    user_input = "Stream AI response with logging"
    logger.info(f"User input: {user_input}")

    stream = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": user_input}],
        stream=True
    )

    collected = []
    async for chunk in stream:
        delta = chunk.choices[0].delta.content or ""
        logger.info(f"Received chunk: {delta}")
        collected.append(delta)

    full_response = ''.join(collected)
    logger.info(f"Full streamed response: {full_response}")

asyncio.run(async_stream_log())

output

2026-04-27 12:00:03,000 - INFO - User input: Stream AI response with logging
2026-04-27 12:00:03,100 - INFO - Received chunk: AI
2026-04-27 12:00:03,200 - INFO - Received chunk:  workflow
2026-04-27 12:00:03,300 - INFO - Received chunk:  logging involves capturing data in real-time.
2026-04-27 12:00:03,400 - INFO - Full streamed response: AI workflow logging involves capturing data in real-time.

Troubleshooting logging issues

If logs do not appear, ensure logging.basicConfig is called before any logging.
Check environment variable OPENAI_API_KEY is set correctly.
For async logging, confirm your event loop is running and await is used properly.
Use logger.error to capture exceptions and stack traces for debugging.

✅

Key Takeaways

Use Python's built-in logging module to capture AI workflow inputs, outputs, and errors.
Wrap AI SDK calls with logging statements for traceability and debugging.
For streaming or async workflows, log each chunk or event as it arrives.
Always configure logging before running AI calls to ensure logs are captured.
Check environment variables and exception logs to troubleshoot issues.

Verified 2026-04 · gpt-4o, gpt-4o-mini

Verify ↗