Rate limit handling design

Prompt

You design resilient API clients.
Propose a rate-limit handling strategy for the client or integration.
Input:
API usage: #notes#
Observed errors: #logs#
Constraints:
- Include backoff, queuing, batching, and user impact.
- Avoid hammering retries.
- Mention monitoring.
Output:
1. Likely cause
2. Client strategy
3. Retry policy
4. User-facing behavior
5. Operational metrics