--- title: "Dashboard KPI Reference" source: /gc-surge/admin/dashboard-kpi-reference locale: en updated: 2026-09-23 --- ## Summary This reference explains the Alarm Center metrics used to assess processing speed, alarm throughput, site coverage, and device coverage. Read related metrics together; no single KPI is a complete measure of Operator performance. Covers: Alarm Center KPI cards, Operator KPI cards, Avg. Operator Processing Time, Top Performers, Picked sites, Alarms handled, Unique devices, How to interpret the metrics together, Related operational panels. ## Alarm Center KPI cards The Admin dashboard shows four live cards for the selected Viewing Period: - **Operator-handled alarms** — alarms brought to closure by Operators. - **Filtered by NOVA99x** — alarms NOVA99x classified as non-actionable and removed before they reached any Operator. A high share means most noise was filtered automatically. Filtered or not, every alarm still counts toward billing. - **Avg. Operator Processing Time** — team-wide average handling time. - **Avg. Cameras per Operator** — average live camera workload per active Operator. ![image-20260720-121134.png](/api/media/file/image-20260720-121134.png) ## Operator KPI cards An Operator sees a different set of four cards, scoped to their own shift rather than the team. The info icon on each card opens a tooltip with the definition and the min and max behind the figure. - **SITES ASSIGNED** — sites assigned to that Operator during the shift. - **CAMERA ASSIGNED** — cameras currently assigned to them for alarm handling. - **ALARM PROCESSED** — alarms they have closed during the shift. - **ALARM PROCESSING** — their own average handling time in seconds, the personal counterpart to the team-wide Avg. Operator Processing Time above. The four cards follow the selected **Viewing Period** rather than the shift alone, so on the default *Last 24 hours* a card can already show a figure at the moment a shift starts. See [Alarm Center - Operator View](/operator/alarm-center-operator-view). ![image-20260720-121235.png](/api/media/file/image-20260720-121235.png) ## Avg. Operator Processing Time Avg. Operator Processing Time is the team-wide average time from an Operator taking responsibility for a site or alarm to closing its alarms. Use these ranges as investigation signals, not automatic pass/fail rules: - **About 30–90 seconds per alarm** — generally healthy. - **Under 15 seconds** — may indicate alarms are being bulk-closed or closed without enough investigation. - **Over 3 minutes** — may indicate difficult events, workflow friction, inadequate training, or a technical problem. Always compare this KPI with Alarms Handled, closure tags, and the complexity of the assigned sites. ## Top Performers Top Performers is subtitled *Track operator throughput, processing time, and device coverage*. Three summary chips sit in its header: **Operators** (how many appear), **Top alarms** (the leader’s alarms-handled figure), and **Leader** (their name). A **Search** box filters the list, with a *Showing x-y of n* counter beside it. The table contains **Operator**, **Picked sites**, **Avg. processing (sec)**, **Alarms handled**, and **Unique devices**; rows carry a rank badge (#1, #2, …) and are sorted by **Alarms handled** descending by default. ## Picked sites Picked sites counts how many distinct sites the Operator handled events from. - Higher values usually mean broader coverage. - Consistently low values may mean the Operator is stuck on assigned sites or skipping cross-team coverage. - Low values can also be correct for dedicated-customer or dedicated-site roles. - Compare Picked sites with Unique devices: one site may contain many devices. ## Alarms handled Alarms handled counts alarms the Operator brought to a terminal state: closed with a closure tag. Alarms that were opened but not taken to a terminal state do not count. This is a throughput metric, not the number of alarms that appeared. A high count can indicate strong throughput, but it can also reflect bulk-closing. Review Avg. Processing Time and closure-tag distribution before drawing a conclusion. ## Unique devices Unique devices counts the distinct cameras or devices whose events the Operator handled. Use it with Picked sites to understand workload density: - Few sites and many devices — concentrated, device-dense coverage. - Many sites and few devices — broad but lighter coverage. - Similar picked-site counts can represent very different workloads when device counts differ. ## How to interpret the metrics together Use combinations rather than isolated values: - **High Alarms Handled + healthy processing time** — likely strong throughput. - **High Alarms Handled + under-15-second processing** — review for bulk-closing or insufficient investigation. - **Low Alarms Handled + over-3-minute processing** — check for difficult alarms, training needs, or workflow friction. - **Low Picked sites + high Unique devices** — may be a legitimately dense assignment. - **Low Picked sites + low Unique devices** — check whether the Operator is stuck or avoiding shared coverage. ## Related operational panels The Admin dashboard also includes: - **Sites Unassigned to Operators** — uncovered sites with active alarms. - **Sites with Assigned Operators** — current site ownership. - **Rollout Progress** — site onboarding and device connectivity. - **Users** — active shift coverage and user presence. For how to act on those panels, see [Alarm Center - Admin View](/admin/alarm-center-admin-view).