Conversation
Radar replaces the Resources Hub with the gaps a fleet manager has to decide about, one item per record: maintenance and inspection schedules due, overdue or blocked work orders, open issues, failed inspections with no follow-up, unresolved inspection follow-ups, stale drafts, inspection links expiring unused, vehicles flagged inspection_failed, drivers late for a shift, on shift without a vehicle, or ending a shift with orders still assigned, licences and leases expiring, unmatched fuel, low stock, and the notices people write themselves. RadarRules is pure: it takes plain rows and a clock and answers with items, pill counts and a summary, so every rule is tested against a fixture and a fixed morning. RadarItemState is the one piece that touches the database: acknowledge, snooze, assign and plan live on the core alerts table, a row created only when someone acts, found again by the item key, and resolved automatically when the record's gap closes. RadarController loads each source in its own try/catch so one failing table degrades the page rather than blanking it. Three fixes the rules exposed: a driver's current shift filtered on a status the schedule_items enum does not have; a part had three different definitions of "low stock"; an inspection schedule produced a preventive maintenance work order.
The brief is arithmetic, not a black box: each category row scores its open gaps by severity (capped per rule), the overall score is the mean, and yesterday's score comes from a per-company setting so the delta is real. Three or four sentences name the records behind the morning with links, and the decision cards each describe one call the console can make in a click (raise the work order for a failed inspection, open a work order for an overdue schedule, put an idle vehicle on a driver without one, match a fuel transaction to the vehicle its card belongs to), with the undo where there is one. The agenda answers the same items placed on the next 24 hours or 7 days: an overdue band left of now, four lanes, a later rail, and an anytime tray for undated work, which a drop gives a time via `planned_at`. Every live shift becomes a bar on the shifts lane; a shift ending with orders still assigned gets a handover card naming the nearest on-shift driver with capacity, and an endpoint extends a shift by the hour. Also: items can be narrowed to one category or to the caller's own assignments, and driver subjects carry a phone for the Call action.
The hub was a landing page of counts and links. Radar is the triage list behind them: one row per gap, grouped by due date, with count pills as the only filter, Open / Snoozed / Resolved tabs, and saved views kept per user. A row's hover rail carries the record's own action (raise the work order, assign the vehicle, match the transaction, send the PIN) beside acknowledge, snooze and assign; a drawer shows the failed inspection items or the shift behind a row. Selection turns the rail into a bulk bar, and J/K/X/E/S/A/Enter work the list from the keyboard. Above the list the morning brief scores each category with visible arithmetic, writes a few linked sentences about the day, and offers one decision per card, each a single call with an undo strip. The Agenda toggle lays the same items on the next 24 hours or 7 days: an overdue band, four lanes, a later rail and an anytime tray a drop can schedule from, with every live shift as a bar and a handover card naming the nearest driver with capacity to take a leaving driver's orders. Styling stays on the console's palette: sky-500 for the active pill and primary actions, the badge palette for chips and severities, the hub's rose and amber tones for overdue and today, and dark mode through the same body[data-theme] overrides as the rest of the engine.
"Radar · 3 overdue" as a KPI tile that links to the page, registered with the other tiles and as a Fleet-Ops header shortcut. Two things the test build needed: the dummy app now loads only en-us translations, because ember-intl hydrates every locale at boot and Chrome 152 ships no Intl data for `mn`, which took every rendering test down before it began; and the sidebar's operations-monitor component gets the app re-export every engine component needs to resolve outside the engine.
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The Resources Hub was a landing page of counts and links. Radar is the triage list behind them, built from the "Fleet Health Concepts" design review: Ops Desk (inbox) as the page, the Morning Brief as a strip on top, and Agenda/Handover as a view toggle, on the console's own palette and dark-mode conventions.
J K X E S A ⏎work the list from the keyboard.inspection_failed, inspection-type schedules due.Items are computed live from existing records (
RadarRules, pure and fixture-tested); state (acknowledge, snooze, assign, plan) lives on the corealertstable and auto-resolves when a gap closes. Each source loads in its own try/catch so one failing table degrades the page instead of blanking it.Three bugs the rules exposed are fixed:
Driver::currentShift()filtered on a status theschedule_itemsenum does not have, parts had three different definitions of "low stock", and inspection schedules producedpreventive_maintenancework orders.Depends on
fleetbase/core-api#258 (branch
feature/alerts-snooze-assignment, targetingrelease/v1.6.62; addssnoozed_until,snoozed_by_uuid,assigned_to_uuid,planned_attoalerts, plusunsnooze(),assignTo(),snoozed()/active()scopes, andacknowledge()now setsstatus = acknowledged). Land and migrate that first.Test plan
RadarRulesTest,RadarItemStateTest,RadarBriefingTest,RadarAgendaTest,RadarControllerTest,RadarRoutesTest(41 tests), plus the existing hub, command, driver-scheduling, model-accessor and route-registration suitesnpm run lint(js, hbs, css, intl)octane:reload, open/fleet-ops/manage; walk pills, tabs, saved views, rail actions, keyboard, bulk, drawer, brief decisions, agenda drag and handover card; dark mode; widget on the dashboardNotes for reviewers: the dummy test app now loads only
en-ustranslations (tests/dummy/config/ember-intl.js) because Chrome 152 ships no Intl data formnand ember-intl hydrates every locale at boot, which took every rendering test down; the sidebar'soperations-monitorcomponent gained theapp/re-export it was missing. The one remaining red test in the hub-index unit module (a devices controller injectinghostRouter) is pre-existing.