SLA & support channels
Service availability
Routic is designed for high availability, backed by automatic failover and retry mechanisms:
| Mechanism | Description |
|---|---|
| Gateway auto-retry | When a model provider fails, the gateway automatically retries up to 3 times. |
| Routing aliases | Routes like auto/chat can fall back to alternative model providers when enabled for your account. |
| Health monitoring | API and gateway endpoints are monitored at 60-second intervals. |
What to expect
- API availability target: 99.9% monthly uptime (excluding scheduled maintenance).
- Incident response: Critical issues are addressed within 4 hours during business hours.
- Maintenance windows: Scheduled maintenance is communicated at least 24 hours in advance.
Support channels
| Channel | Use for | Response time |
|---|---|---|
| support@routic.ai | API errors, integration help, debugging with request_id | 1–3 business days |
| contact@routic.ai | Billing questions, invoice requests, account issues, general inquiries | 1–2 business days |
| Urgent issues | Service outages, data integrity concerns | Within 4 hours (business hours) |
For urgent issues, email support@routic.ai with URGENT in the subject line.
How to contact support
When contacting support, include the following to get a faster response:
- Your account ID — found in your dashboard.
request_id— for API-related issues, always include the request ID from the error response. See Observability & request IDs.- Timestamp — when the issue occurred, in UTC or your local timezone.
- Description — what you expected vs. what happened.
- Error response — the full JSON error body, if available.
Reporting bugs
If you find a bug in the API:
- Check the HTTP & errors page to confirm it's not expected behavior.
- Verify you're using the latest version of the OpenAI SDK (or your HTTP client).
- Create a minimal reproduction and send it to support with the
request_id.
Feature requests
Feature requests are welcome. Contact support or your account manager with:
- The problem you're trying to solve
- The current workaround (if any)
- How the feature would help
Rate limit increases
If you need higher rate limits, see Rate limits → Requesting higher limits.