Anyone else seeing a massive uptick in API errors from Leonardo over the last 48 hours? My CI pipeline for generating product thumbnails has been failing consistently. Their status page is all green, but the reality is far from it.
Here's what I'm getting:
* **429s** that don't align with my documented rate limits. My traffic hasn't spiked.
* Random **502 Bad Gateway** errors on image generation endpoints.
* **Latency spikes** from the usual ~1.8s to 8s+, then timeouts.
I ran a simple script to poll their `/models` endpoint every 5 minutes. The availability dropped to 87% for me yesterday. This isn't a "blip."
```python
import requests, time
from datetime import datetime
endpoint = "https://cloud.leonardo.ai/api/rest/v1/models"
headers = {"Authorization": "Bearer YOUR_KEY"}
success = 0; total = 0
for _ in range(288): # 24 hours
try:
resp = requests.get(endpoint, headers=headers, timeout=10)
if resp.status_code == 200:
success += 1
except:
pass
total += 1
time.sleep(300)
print(f"Availability: {(success/total)*100:.2f}%")
```
Are they doing unannounced infra changes or is load just overwhelming them? For a paid service, this level of instability is unacceptable for production use. What specific errors are you all hitting?
-- bb
-- bb