5xx Server Errors

{
  "error": {
    "code": "upstream_error",
    "message": "Upstream provider returned 503. The provider may be experiencing downtime.",
    "request_id": "req_01hwz3kj4p5qm8n9v2t6yt",
    "source": "provider",
    "provider": "openai",
    "provider_status": 503,
    "retryable": true
  }
}

# Check provider status pages
# OpenAI: https://status.openai.com
# Anthropic: https://status.anthropic.com

{
  "error": {
    "code": "gateway_error",
    "message": "An internal error occurred. Our team has been notified.",
    "request_id": "req_01hwz3kj4p5qm8n9v2t6yu",
    "source": "curate-me",
    "retryable": true
  }
}

{
  "error": {
    "code": "cost_recorder_error",
    "message": "Failed to record request cost. The request was NOT forwarded to the provider.",
    "request_id": "req_01hwz3kj4p5qm8n9v2t6yv",
    "source": "curate-me",
    "retryable": true
  }
}

{
  "error": {
    "code": "proxy_timeout",
    "message": "Request to provider timed out after 120s.",
    "request_id": "req_01hwz3kj4p5qm8n9v2t6yw",
    "source": "provider",
    "retryable": true
  }
}

import time
from openai import OpenAI, APIStatusError
 
client = OpenAI(
    base_url="https://api.curate-me.ai/v1/openai",
    api_key="cm_sk_your_key",
    max_retries=3,  # Handles 5xx automatically
)
 
# Or manual with exponential backoff
def call_with_exponential_backoff(messages, max_attempts=4):
    delay = 1.0
    for attempt in range(max_attempts):
        try:
            return client.chat.completions.create(
                model="gpt-4o",
                messages=messages,
            )
        except APIStatusError as e:
            if e.status_code >= 500 and attempt < max_attempts - 1:
                print(f"5xx error (attempt {attempt + 1}). Retrying in {delay:.1f}s...")
                time.sleep(delay)
                delay = min(delay * 2, 30)  # cap at 30s
            else:
                raise