Dashboard
/dashboard — what is waiting, and what is late.
Everything on the page is scoped to one laboratory and one day.
What it shows
| Tile | Answers |
|---|---|
| Specimens received today | How much work arrived |
| Unacknowledged criticals | What is waiting on a person |
| Median receipt-to-result | How long the laboratory is actually taking |
| Receipt-to-result, by day | The same figure over time |
| Department load | Where the work is sitting |
| Oldest unreleased results | What has been waiting longest, linking straight to review |
Every figure is computed in the database
Not one number on this page is calculated in the browser. Each is a bounded, tenant-filtered query, which is what makes the tiles, the department bars, the table and the chart unable to disagree with each other.
A figure it cannot compute says so
The median turnaround tile refuses a number it cannot produce, and "not measured" is its own separate question — never inferred from an empty chart. "No releases this week" and "not enough history for a median" are different statements, and showing one when the other is true is how a dashboard misleads quietly.
This rule has a specific origin: the application LabFlow replaces produced that figure with a random function behind a swallowed permission error, and the result was indistinguishable from a real one.
For the same reason there is no revenue tile — there is no billing data behind this deployment, and a currency total is the number a dashboard is most often asked to invent — and nothing here forecasts. Where arithmetic is used it is described as arithmetic.
Unacknowledged criticals
New ones appear the moment they are flagged, and they escalate whether or not this page is open. The dashboard reports the queue; it is not the mechanism.