Performance

Traffic patterns and heatmap

The traffic patterns card helps you understand when your fleet is busiest. A date-by-hour heatmap shows call density across the selected time range, while breakdowns by proxy type and call route reveal how traffic flows through the platform.

Heatmap visualization

The heatmap is a sparse matrix with dates on the vertical axis and hours (0–23) on the horizontal axis. Each cell represents one hour of one day. Cell color intensity reflects call volume — darker cells indicate more calls. Hover over any cell to see the exact count.

The heatmap applies your local timezone offset so that hours align with your wall-clock time. All timestamps stored in the database are UTC; the conversion happens at render time.

Reading the heatmap

The heatmap uses a sequential color scale from light to dark:

Empty / lightest — zero or very few calls in that hour
Medium intensity — moderate traffic relative to the range peak
Darkest — the highest-traffic hours in the selected range

Look for horizontal bands (consistent busy hours across days) to identify daily traffic patterns. Vertical bands (all hours busy on one day) may indicate batch processing or a traffic event.

Timezone handling

Hours on the horizontal axis are in your local timezone, detected from your browser. If your fleet serves users across timezones, remember that the heatmap reflects your perspective — a cluster at hour 14 means 2 PM your time, which may be a different wall-clock hour for end users.

Proxy type breakdown

Below the heatmap, the proxy type breakdown shows how calls are distributed across the four proxy types: A2A, MCP, ACP, and REST. Each type is shown with its count and a proportional bar so you can see which protocols carry the most traffic.

Call route breakdown

The call route breakdown categorizes calls by their routing path — direct connections, scheduled calls, pool-routed calls, and broadcast-routed calls. This tells you how much of your traffic uses standard point-to-point connections versus advanced routing features.

Agent-specific vs fleet-wide view

When the agent filter is set to a specific agent, the heatmap and breakdowns show only that agent's traffic. Switch to "All agents" for a fleet-wide view. The fleet view aggregates all agents, which is useful for capacity planning and identifying system-wide patterns.

Use the agent-specific view to compare traffic patterns between agents. If one agent has a completely different heatmap shape, it may serve a different user base or handle batch jobs on a different schedule.

Monitor schedule, fallback, and pool health metrics. See Routing health →