Skip to content

Dashboard

The dashboard is the landing page of the Gatewyse admin panel. It provides a real-time operational snapshot of your gateway, surfacing key metrics, provider health, usage trends, and recent request activity.

Metrics Cards

The top of the dashboard displays four summary cards, each with a sparkline trend indicator:

MetricFormatDescription
Total RequestsNumberTotal gateway requests in the last 24 hours
Active UsersNumberUsers currently making requests
Cost TodayCurrencyEstimated spend for the current day
Cache Hit RatePercentageSemantic cache effectiveness

Each card shows a trend arrow (positive or negative percentage) so you can spot changes at a glance.

Provider Health Map

Below the metrics cards, the Provider Health Map renders a status tile for every configured provider. Each tile shows:

  • Provider name and type (e.g. OpenAI, Anthropic, Google)
  • Health status color-coded as one of:
    • healthy — provider responding normally
    • degraded — elevated latency or error rate
    • down — provider unreachable
    • unknown — no health check data yet

Health data is fetched from the /api/admin/providers/health endpoint and refreshed when the dashboard loads.

Usage Chart

The Request Volume chart displays request counts over time. You can toggle the period between:

  • Hourly — last 24 data points (one per hour)
  • Daily — last 30 days
  • Weekly — last 12 weeks

The chart updates in place when you change the period selector.

Cost Breakdown

The Cost Breakdown panel lists per-provider spending. Each row includes:

  • Provider name
  • Total cost in the current period
  • Number of requests attributed to that provider

This helps you identify which providers are driving your spend.

Recent Requests

The Recent Requests table shows the most recent gateway calls with the following columns:

ColumnDescription
TimestampWhen the request was made
ModelThe model that handled the request
ProviderWhich provider served the request
Statussuccess or error
LatencyRound-trip time in milliseconds
TokensTotal tokens consumed

Error rows are highlighted so you can quickly identify failures.

Data Sources

The dashboard pulls data from two Pinia stores:

  • MetricsStore — fetches real-time metrics (totalRequests, activeUsers, costToday, cacheHitRate, requestsPerMinute, avgLatency, errorRate) and historical time-series data.
  • ProviderStore — fetches the provider list and health status for the health map.

Both stores are loaded in parallel on page mount to minimize initial load time.