Service Outage Protocol
The flow chart described the recommended approach to handling when the service is unavailable or a device is offline.
Timeout-Retry Sequence
API latency is < 1s in available regions (excluding any network dependency)
We suggest setting:
- Timeout = 2.5s
This will happen rarely and will be a good indication of a fault.
Offline mode
If the API fails/timeout more than twice in a 15min period, enter “Offline Mode”.
- Cease API calls.
- Report error – on screen without preventing usage and/or notify staff.
- Default to lookup menu for usage.
Fast check for server status
The health check endpoint, available at the API index, quickly confirms if the server is online and responsive. Since no inference is performed at this endpoint, it provides a fast response.
Updated 11 months ago