Read the Live field

Use the Live field to see, in one picture, how the agents of a zone behaved over the last 30 days: which acted within policy, which strained a dimension, and which were stopped at the boundary. Anyone who can read a zone can read its Live field.

Open the Autonomy Zone page and select a unit. The Live field sits on the zone’s overview, under the zone at a glance. It always covers the last 30 days; there is no window to choose. Manage Autonomy Zones explains the rest of the page.

Read the court

The field is drawn as a basketball court. Every line and every area of the court has one meaning, and the court itself never carries a status color: only the agents do.

Part of the court	What it means
Center circle	Agents with no anomaly in 30 days. Its heading reads WITHIN POLICY, or NO INTENT POLICIES when the zone has no live intent policy to evaluate
Ring, between the center circle and the three-point line	Agents with 1 to 4 anomalies in 30 days
Outer band, between the three-point line and the sideline	Agents with 5 or more anomalies in 30 days. Its tint fades to the center’s when no agent has that many
Sideline	The policy boundary. Only an agent with a Block in 30 days sits on it, on the stretch of the dimension it was blocked on most. A red bar just outside the sideline spans those agents
Lane lines	Split the court into five territories, one per dimension: Security at the top, Economics on the right, Quality on the left, Safety at the bottom left, Compliance at the bottom right
BENCH, under the court on the left	Agents with no activity in 30 days
NOT MONITORED, under the court on the right	Known agents: they have no adapter, so their actions are not visible and no anomaly can be counted

An anomaly is an action that violates an intent policy. The court counts them per agent over the last 30 days.

Read an agent

Each agent is one dot. Every dot has the same size and halo; only its fill changes with the agent’s anomaly load:

Dot	Anomaly load in 30 days
Teal	No anomaly (gray when the zone has no live intent policy, since nothing was evaluated)
Amber	1 to 4 anomalies
Red	5 or more anomalies

Red reads the agent’s anomaly load. It never says that the agent was blocked: an agent’s Block shows only by its place on the sideline.

Where a dot sits tells you the rest:

• Distance from the center grows with the number of anomalies, so an agent with many anomalies sits farther out than one with a few.
• Territory is the dimension the agent hit most. When two dimensions tie, the agent sits in the one of its latest anomaly, and its hover card says so. An agent with a Block sits in the territory of the dimension it was blocked on.
• Notches, small marks on the rim of a dot, show the other dimensions the agent also had anomalies on. Each points at that dimension’s territory.

Agents with no anomaly sit on a fixed pattern inside the center circle. The same data always draws the same court: nothing moves at random, and an agent keeps its seat from one visit to the next.

Hover over a dot, or focus it with the keyboard, to open its card: its unit and when it was last seen, its anomalies per dimension, its Blocks, and its decisions in 30 days by verdict. Use the arrow keys to step through the agents, most anomalies first.

Read the sidelines

Each stretch of the sideline belongs to one dimension and shows the intent policies the zone applies to it:

Sideline	Meaning
Thick	Policies in Enforce: the more there are, the thicker the line
Thin	Monitor only
Dashed, captioned No policies	No intent policy on that dimension
Dotted line just inside	The dimension has both inherited policies and policies added in this zone

A quiet territory behind a dashed sideline is quiet because nothing evaluates it, not because its agents are safe.

Watch the decisions

Trails replay the decisions OpenLatch recorded in the last 30 days. Each trail runs from an agent to the sideline of the decision’s dimension, colored by its verdict: Allow, Ask or Block. Optimize decisions are counted in the summary and in the agent’s card, and never drawn as a trail.

Nothing is drawn that was not recorded: an agent with no decision has no trail, and a zone with no decision in 30 days shows no motion at all. The summary above the court reads the totals, for example 12 decisions: 8 Allow · 2 Ask · 1 Block · 1 Optimize.

Choose Pause to stop the replay. When the replay is paused, or when your system asks for reduced motion, the 20 most recent decisions show as static ticks along the sidelines instead.

Filter, zoom and explore

Control	What it does
All	Every agent of the zone, on and off the court
Active	Agents seen in the last 30 days that OpenLatch can observe. Hides the bench and Not monitored
Anomalies	Agents with at least one anomaly in 30 days. Hides the bench and Not monitored
Legend	What each line, band and mark means
Dimensions	Per dimension: its intent policies by mode, and its anomalies and agents over 30 days
Zoom	Zoom in up to four times and drag to pan; on a narrow screen the zoom control is hidden

Hover anywhere in a territory, or focus its label, to highlight that dimension: its agents stand out and the rest fade. Each territory label reads its own numbers, such as how many agents, anomalies and Blocks it holds.

Large fleets

The court stays readable however many agents a zone holds:

• At most ten agents carry a name on the court. Agents with anomalies that have no name on the court are counted in the summary as unlabeled; choose that count to list them, most anomalies first.
• When a territory has more agents with 5 or more anomalies than its strip can draw one by one, the strip shows a red dot texture with the exact count. Hover over it to list those agents. Agents with a Block always stay as dots on the sideline.
• When more than 250 agents have no anomaly, the center circle shows a teal texture instead of dots; hover over it for the count.
• The court draws up to 500 agents. When the zone holds more, the summary says how many of the total are shown.