Status
MCP service status
The real-time operational status of the Trinity Insights MCP Server is published on the canonical status page trinityinsights.io/status. This page documents the posture, service objectives (SLO) and user-side incident procedure.
Real-time status
The current status of all Trinity components - application, intelligence, infrastructure, MCP - is published and refreshed periodically on:
→ View Trinity Insights statusService objectives (SLO)
Trinity Insights commits to the following qualitative objectives for the MCP Server. These objectives are not a contractual SLA; they represent the founder's quality-of-service commitment.
| Dimension | Qualitative objective |
|---|---|
| Availability | High availability outside planned maintenance windows |
| Latency | Sub-second responses on standard tool calls |
| Error rate | Server errors (5xx) kept at a low level, excluding legitimate client errors (tier denied, rate limit) |
| Catalog freshness | The indicator catalog is synced regularly with the dashboard source of truth |
| Recovery (RTO) | Rapid recovery via automatic rollback when a deployment introduces a regression |
In case of a user-side incident
If you observe abnormal behavior from the MCP Server:
- Check trinityinsights.io/status to see whether an incident is in progress
- Confirm that your API key has not been revoked (Dashboard → Settings → MCP API Keys)
- Confirm that your dashboard tier is active (Dashboard → Settings → Subscription)
- Inspect the logs of your local MCP client (Claude Desktop, Cursor, etc.)
- If the issue persists, contact [email protected] with a timestamp and a description of the symptom
Maintenance windows
Deployments follow a standard Trinity DR procedure with automatic rollback in case of regression. Major maintenance windows are announced by email to Pro, Performance and Team users at least 48 hours in advance. No impactful maintenance is scheduled on a regular cadence.