Coverage

Coverage tells you how well the model describes each asset. It is computed on every read and save and never stored, because history and alarm setpoints change without anyone saving the model. What a read sees of the tags (history, setpoint alarms, whether each tag and alarm exists) comes from one read of each tag provider the model names and is kept for 60 seconds per collection, keyed by the model's revision and dropped by a save: a tag edit shows within a minute, or at once after a save. Per asset, each score 0–1 or not applicable: tags (has at least one), units (share of measurement tags with a unit), ranges (share with a normal range or a limit; an alarm setpoint on the tag counts), relations (in at least one relation), rules, documents (at least one existing document), history (share of tags with history enabled) and state (1 when the asset names a run signal, else 0; not applicable until some asset of the collection names a state, so a model that does not use Anomaly Watch keeps its score). A UDT member counts the history and setpoint alarms it inherits from its type. The asset's score is the mean of what applies; the collection's the mean of its assets'. The editor's to-do lists the assets below full coverage, ordered by their open findings and then by their alarm activations in the last 7 days.