At a glance
Share of calls to the Japan Post API that returned an error (4xx or 5xx) over the trailing 24 hours. It is the integration-health pulse for the connector: when it climbs, label generation, tracking lookups, and claim queries are failing somewhere between Vortex IQ and Japan Post. The alert fires above 2 percent, which for a busy despatch operation is the line between background noise and a real fault.
What it tracks
The card computesfailed_calls ÷ total_calls across all Japan Post API endpoints (label and manifest generation, tracking, claims, billing) over the rolling 24-hour window, expressed as a percentage. A sustained reading above the >2% threshold usually has one of three causes: an expired or revoked auth token (4xx, cross-check Days to Token Expiry), a Japan Post infrastructure outage (5xx, cross-check Japan Post Tracking API Unavailable / 5xx), or malformed request payloads such as bad address or weight data. A 24-hour window is short enough to catch an outage quickly but long enough to smooth single-minute blips. The downstream symptom shows on Label Generation Success: when error rate spikes, label success drops in lock-step. Use the error class to decide the action: a 4xx storm is usually yours to fix (token, payload), a 5xx storm is Japan Post’s to fix and the right move is to wait and retry rather than re-authenticate.