Skip to content

Readouts and boards

This page documents the readouts and summary boards you see at each scope in a grid: the header, the graph root, and an individual service platform.

Header readouts (persistent)

The header runs across the top of the screen.

Readout Position Meaning
Grid name Top center The current grid (for example DEMO-TRACEGEN-PHANTOM-EASTUS-PROD)
Current location Top center, directly beneath the grid name Where you are right now (for example "Portal Room"). This is your "where am I" readout
% Optimal Top left Tessa's energy state. Not system health. See Tessa commands and runtime for the energy-state ladder
XP (e.g. "0 / 10000") Top left, below % Optimal Cosmetic experience accumulated from using the app. Unfinished
Roaming ON/OFF Top right Toggles Mosey autopilot / AFK mode
Pro Tips ON/OFF Top right Toggles placards on or off globally
Clock Top right Current time
Play/pause Top center, below the location Transport control; see Navigation model

Graph root boards (whole-grid scope)

Entered via the SW portal from the Core, or by teleporting down to the graph. These boards describe the entire grid, not any single service.

Recent Errors

Column Description
Service Service that produced the error
Instance Instance identifier
Time Timestamp
Operation The operation that errored
Duration How long the operation took

Services

Column Description
Service Service name
Instance Instance identifier
Status Red "Error" or green "Active"
Ops Operation count
Last Activity Time since last activity
Connections Connection count

Dependencies

Column Description
Dependency Dependency name
Type The dependency's type. "Phantom" is a first-class Type value: an expected dependency inferred from callers but not itself instrumented and not reporting
Status Health status
Ops Operation count
Last Activity Time since last activity

Services are instrumented and emit telemetry directly. Dependencies are known indirectly, from the telemetry of the services that call them. See Rendering absence for why absence is rendered this way.

LLM Costs

Column Description
Service Service name
Cost 60s Cost over the last 60 seconds
Cost/s Cost per second
Tokens 60s Tokens consumed over the last 60 seconds
Input/Output Input vs. output token split

Service platform boards (single-service scope)

Reached by teleporting to a service node, either by line of sight or by clicking its row on a graph-root board. These boards are scoped to the one service you are standing on.

Board Columns / content
Recent Errors This service's error feed
Recent Logs This service's log lines
Calling Services Inbound callers of this service, with status, ops, and last activity
LLM Cost Usage / Input Tokens / Output Tokens / Cost / Cost/s, for this service

The platform's neon (edge lighting) is red or green for the health of the service you are standing on. See Color legend.

Tessa's conversational context locks to the service platform you are standing on.

Tooltips

Most objects, platforms, and cubes carry a hover tooltip giving basic information about what they are. Hover an object to see its tooltip, and press F6 to copy the tooltip contents to the clipboard.

Hovering a service node shows that service's tooltip:

Field group Fields
Status Operations, Last Activity, Connections
Performance (last 60s) Throughput (ops/s and total), Latency p50 / p95 / p99, Errors