Feature Reference

The per-project command center (sidebar → Network Operation Center → project → NOC Dashboard). It shows:

• KPI strip — total sites, online, offline, online %, unique users (30d), and traffic (30d).
• Outage Triage — live down-sites, clustered (see below).
• Sites Map — site markers color-coded by status, plus a list of sites with no stored location.
• Uptime & Data Usage charts — daily online % and traffic trends.
• API Error Logs — recent vendor API failures (deduplicated), useful for telling a real outage apart from a data-collection gap.

Route: /noc/{projectId}. See the Routes & URLs reference for the full list.

A once-per-day briefing that is frozen when first generated, so everyone sees the same morning snapshot. It includes a health verdict (healthy / degraded / major outage), a narrative lead, highlight chips with prior-day deltas, and an outage board (area-wide clusters, isolated faults, recovered overnight, newly-down). Use the date picker to scroll back through saved reports.

The report text is frozen for the day, but the embedded map is always live — on past dates it reflects current status, not that day's. Route: /noc/{projectId}/report.

Outage Triage clusters currently-down sites and suggests the next action. It's hybrid: the clustering is deterministic geometry (sites within ~10 km, three or more = an area-wide cluster), and an optional Claude layer writes the dispatch briefing on top.

If the AI is unavailable it degrades gracefully to the heuristic output — it never blocks. Either way, the recommended action always follows the escalation ladder: remote reset → call contact → dispatch only as a last resort.

On every site page, under Uptime and Alerts. The daily uptime figure can't say when an access point was down — two of four APs up reads as 50% whether that was a weekend shutdown, a six-day outage or a site that keeps flapping. This card shows each AP's poll for every hour.

• A chip per AP — its state now and since when, e.g. “AP2 down 1d 16h since Fri 25 Sep, 10:00”. “Now” is the header's: the APs' status this minute, which can be newer than the last hourly poll. The Now cell at the end of each strip shows it too.
• One rule for up and down — if the system says an AP is up, it is up. A dark site has been down since any AP was last seen up. The Sites list, map, digest, tickets, Planning and the crew app all use this rule.
• One strip per AP plus an APs up bar across the top; hover to read the exact hour. 1d, 7d, 14d or 30d (default 30d).
• Down 6 hours or more — every stretch an AP was down that long, with start, end, duration and “ongoing”. These lines are what a DICT justification needs.
• Grey means no reading — the collector missed that hour. It is never counted as down.

The Uptime chart beside it is built from the same polls: each day is the share of hours its APs were up, and “Outages” counts the times every AP went down — so the header, the chart and the strips always agree.

The same data, as a day-by-hour grid, is in the View panel of DICT escalation.

The geographic view of sites across a project (and the cross-project Maps view for admins). Marker color reflects status, so a cluster of red markers in one area is an immediate signal of a shared power or backhaul cause rather than independent site faults. Route: /maps.

The Reports hub produces exportable summaries:

• Downtime — device/group downtime over a date range (/reports/downtime).
• Users — unique users and data usage per site (/reports/users).
• Billing — traffic/billing summaries (/reports/billing).

Each project has a helpdesk view where outages become tickets with severity, SLA timers (acknowledge / respond / resolve), and escalation level. Isolated faults and area-wide clusters can each generate a ticket, which is how a Coordinator's escalation gets tracked through to resolution. Route: /noc/{projectId}/helpdesk.