Azure OpenAI error codes reference
429 for RateLimitError, 401 for authentication issues, and 400 for bad requests. Use the AzureOpenAI client with proper error handling to catch and respond to these codes.api_error RateLimitError automatically.Why this happens
Azure OpenAI API errors occur due to invalid requests, authentication failures, or exceeding usage limits. For example, a 429 RateLimitError happens when your app sends too many requests too quickly. A 401 Unauthorized error indicates missing or invalid API keys. A 400 Bad Request error usually means malformed input or unsupported parameters.
Typical error output looks like:
{"error": {"code": "RateLimitExceeded", "message": "You have exceeded your quota."}}Example broken code without error handling:
from openai import AzureOpenAI
import os
client = AzureOpenAI(
api_key=os.environ["AZURE_OPENAI_API_KEY"],
azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
api_version="2024-02-01"
)
response = client.chat.completions.create(
model=os.environ["AZURE_OPENAI_DEPLOYMENT"],
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Traceback (most recent call last): ... openai.error.RateLimitError: You have exceeded your quota.
The fix
Wrap your Azure OpenAI API calls with try-except blocks to catch errors like RateLimitError and AuthenticationError. Implement exponential backoff retries to handle transient rate limits. Validate your API key and endpoint environment variables to avoid 401 errors.
This code retries on rate limit errors and prints the response on success:
from openai import AzureOpenAI, RateLimitError, AuthenticationError, OpenAIError
import os
import time
client = AzureOpenAI(
api_key=os.environ["AZURE_OPENAI_API_KEY"],
azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
api_version="2024-02-01"
)
max_retries = 5
for attempt in range(max_retries):
try:
response = client.chat.completions.create(
model=os.environ["AZURE_OPENAI_DEPLOYMENT"],
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)
break
except RateLimitError as e:
wait_time = 2 ** attempt
print(f"Rate limit hit, retrying in {wait_time} seconds...")
time.sleep(wait_time)
except AuthenticationError as e:
print("Authentication failed: check your API key and endpoint.")
break
except OpenAIError as e:
print(f"OpenAI API error: {e}")
breakHello
Preventing it in production
- Use environment variables for API keys and endpoints to avoid misconfiguration.
- Implement exponential backoff retries on
429 RateLimitErrorto gracefully handle throttling. - Validate input data to prevent
400 Bad Requesterrors. - Monitor usage quotas in Azure Portal to avoid unexpected limits.
- Log errors and alert on repeated failures for proactive maintenance.
Key Takeaways
- Always handle
RateLimitErrorwith retries and exponential backoff in Azure OpenAI calls. - Validate and securely store your API keys and endpoints in environment variables to prevent authentication errors.
- Monitor and log API errors to maintain reliable production AI applications.