Status
Every service, from real requests, the last 30 days. Each bar is one UTC day; hover it for the hours with errors.
GET https://api.relayswitch.io/v1/status · generated 2026-10-26T05:14:01.528Z
Last incident Oct 24, 2026, 3:07 PM UTC: elevated errors
Last incident Oct 25, 2026, 4:07 AM UTC: elevated errors
Last incident Oct 21, 2026, 3:07 AM UTC: elevated errors
Models under strain
Models real traffic shows degraded or down in the last 60 minutes, worst first.
None. Every model with traffic in the window answered.
Retry test
Once a day the gateway sends one real request in each format, the way a request is retried after a failure.
| Format | Result |
|---|---|
| Claude, Anthropic format | Answered on Oct 25, 2026, 6:30 AM UTC, 7407 ms. |
| GPT and others, OpenAI format | Answered on Oct 25, 2026, 6:30 AM UTC, 1562 ms. |
What is checked
Claude, GPT and Gemini show what real requests got. Every hour the gateway counts the past hour's requests to each family: answered or failed. A request answered after a retry counts as answered.
API checks that the gateway can build its model catalog, the same list GET /v1/catalog returns, every hour. It passes when the list has models in it.
Each family is also checked every hour with a one-token request. A family with no requests in 30 days shows that check instead. Checks are not billed to any account.
What each state means
- Operational
- In the latest hour, requests were answered.
- Degraded
- In the latest hour, at least one request in five failed.
- Down
- In the latest hour, requests failed and none were answered.
- No data yet
- No requests and no checks for this service in the last 30 days.
During an incident
- A request that fails upstream with a 5xx or 429 is retried for you. Your price stays the same.
- Every response, including errors, carries a request id header. Keep it; it is how we find your request.
- A 429 with a retry-after header says how many seconds to wait before trying again.
- If a request fails and this page shows every service operational, send us the request id.
The same data is public JSON at https://api.relayswitch.io/v1/status: no key, readable from any origin, cached for 60 seconds. Point your own monitoring at it.