Hey folks 👋
Just ran into something that might help others debugging API issues. My team's integration with Anyword's API started throwing 429 (Too Many Requests) errors this morning during our content generation batch jobs. We've been using it smoothly for weeks, so this was a surprise.
I checked our dashboard and saw the error spike. Our setup:
- We're using the `/generate` endpoint for multiple variations.
- Batch jobs run every 2 hours, each job sends ~50 requests in a span of 5 minutes.
- We implemented basic retry logic with exponential backoff.
Here's our current retry snippet (Python):
```python
from tenacity import retry, stop_after_attempt, wait_exponential
@retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=1, min=4, max=10))
def generate_with_retry(prompt):
# ... API call here
```
**Questions for the community:**
- Has anyone hit the rate limits recently? Did they change?
- What’s the actual rate limit? I couldn’t find clear numbers in the docs.
- Are the limits per endpoint, per API key, or per account?
I’m adding more granular metrics to track per-endpoint calls now. If you’ve found a workaround or have official limits to share, please post below! Might save someone a few hours of debugging 😅
Also curious: if you’re monitoring this in production, what’s your alerting strategy for 429s? We’re using Datadog to alert on error rate spikes, but wondering if others have synthetic checks for API health.
Dashboards or it didn't happen.