Data Model¶

Databús uses two fundamentally different storage mechanisms for two different purposes. Understanding the boundary between them is the key to reading any piece of the system.

Two namespaces, two purposes¶

graph LR
    subgraph "Edge-sensed (vehicle:<id>:*)"
        P["vehicle:&lt;id&gt;:position<br/>hash — GPS + motion"]
        O["vehicle:&lt;id&gt;:occupancy<br/>hash — passenger load"]
        M["vehicle:&lt;id&gt;:metadata<br/>hash — descriptor"]
        CR["vehicle:&lt;id&gt;:current_run<br/>string — run_id"]
    end
    subgraph "Server-computed (run:<id>:* / runs:*)"
        R["run:&lt;id&gt;<br/>hash — full assignment"]
        T["run:&lt;id&gt;:trip<br/>hash — GTFS-RT TripDescriptor"]
        S["run:&lt;id&gt;:vehicle_stop_status<br/>hash — stop relationship"]
        C["run:&lt;id&gt;:congestion_level<br/>hash — congestion"]
        U["run:&lt;id&gt;:stop_time_updates<br/>string — JSON array"]
        LS["runs:last_seen:&lt;id&gt;<br/>string — ISO-8601 timestamp"]
        IP["runs:in_progress<br/>set — active run IDs"]
        TR["runs:tracking<br/>set — run IDs with telemetry"]
    end

vehicle:<id>:* keys are written by the MQTT consumer from raw edge signals. They represent what the vehicle reports.

run:<id>:* and runs:* keys are written by the server — lifecycle actions and the progression step. They represent what the server has concluded from those signals.

This separation enforces the single writer per responsibility principle and makes it clear who owns each piece of data.

Pages in this section¶

Start with Redis state keys — it is the canonical reference that all other pages link to.