Connect a UniFi console once and every gateway, switch and access point on its mapped sites is imported, checked every 60 seconds, reported DOWN when the controller stops listing it, and ticketed in the IT Service Management (ITSM) system your team already uses.
UniFi's own dashboards tell a site what is wrong; they do not open a ticket, and they do not tell your NOC. SmartTile reads the console the way the UniFi app does — sites, devices, boot counters, ports, radios and WAN health — and holds itself to one rule: a device counts as present only when the controller lists it in this poll. An AP that drops off the list, a gateway whose controller stops answering, an AP that reboots between two checks — each is recorded as an outage, counted in availability, and ticketed.
Each UniFi site is mapped to a SmartTile location. A multi-site cloud console's devices are attributed to sites by a person on the Map screen — never guessed.
Import preview labels every unit new, imported or offline. You tick what comes in; offline gear is never imported and a unit already in through another connection is matched by MAC, not duplicated.
Offline in the controller, or missing from its list for 5 minutes, is DOWN. The statistics heartbeat is kept separate, so a stale stats reply can never keep a dead AP green.
A boot counter going backwards is a reboot, recorded DOWN→UP with its real duration; 3 in 60 minutes opens a ticket. A controller that stops answering turns its devices UNKNOWN (grey) — a self-hosting gateway goes DOWN.
With a Controller API key: WAN latency and loss per monitor target, ISP and speed test, every switch port (link, PoE, errors, LLDP), every radio and SSID, temperature, storage, site health — 35 days of history.
Per-AP client lists and an API NetFlow view of LAN top talkers (and application mix where the controller runs DPI), plus bandwidth and uptime charts on the device.
When a site swaps an AP, switch or gateway, Map → Replaced hardware hands the existing SmartTile device to the new unit: same id, name, contacts, ticket rules and history. A person confirms every swap; each is logged.
WAN loss, WAN latency, overheating, port errors and half-duplex links can alert and ticket once sustained; Wi-Fi experience, channel utilisation, PoE headroom and end-of-life are advisories. Off by default.
| Connection | What works |
|---|---|
| Controller API key (recommended) | Status, reboots, CPU / memory, clients, uplink throughput, plus full console telemetry and thresholds through the same key. |
| Classic controller login | Everything the classic API gives, including controller report history for bandwidth and uptime charts. |
| Site Manager API key | Device list and status, reboots, firmware waiting, site client counts, console ISP metrics. No per-port or per-radio detail. |
| Cloud connector | The Integration API reached through api.ui.com with the Site Manager key — no VPN or port-forward. |
A device must be offline for 3 consecutive checks (about 6 minutes) before a ticket opens. If half the fleet drops at once, that is a reporting gap — statuses update, tickets hold.
Tickets go through the normal SmartTile path: the device's auto-ticket flag, the location's automation switch and per-location dedup — so every AP behind one dead uplink is one ticket with notes. Rev.io plus any other ITSM system connected for the company.
Every DOWN, UP and reboot is written to state history, so availability and reports count it. A Rev.io ticket is closed once every device on it is confirmed up for 10 minutes; a long outage stays open for the RFO.
/proxy/network/integration/v1/sitessites (Integration API, API key)…/integration/v1/sites/{id}/devicesdevice list and online state…/sites/{id}/devices/{id}/statistics/latestuptime, CPU, memory, clients, uplink/proxy/network/api/s/{site}/stat/devicefull device detail: ports, radios, temperature/proxy/network/api/s/{site}/stat/healthsite WAN / WLAN / LAN / VPN health/api/s/{site}/stat/staclients and top talkers/api/s/{site}/stat/report/{interval}.{scope}history charts — classic login only (a read)api.ui.com /v1/hosts · /v1/sites · /v1/devicesSite Manager fleet listapi.ui.com /v1/isp-metrics/{type}console WAN latency and lossapi.ui.com /v1/connector/consoles/{id}/network/…cloud connector, no VPNEvery call above reads. SmartTile adopts, configures and upgrades nothing on the console.
API keys and passwords are encrypted at rest (AES-256-GCM) and never returned — the form only shows that one is set; leave it blank to keep it.
A customer admin sees and maps only the controllers their company owns; another tenant's controller id answers not found.
The integration, uptime history, ticket confirmation, down and reboot detection, full telemetry and hardware replace each have their own setting. Every replace is in the transaction log.
| Objection | Answer |
|---|---|
| “UniFi already alerts us.” | UniFi e-mails a person. SmartTile records the outage, counts reboots in availability and opens the ticket in your ITSM system — one per site, not one per AP. |
| “A controller hiccup will flood us with tickets.” | Three checks in a row before a ticket, a fleet-share hold, and a blind controller turns devices grey (UNKNOWN), not red. |
| “We swap APs all the time.” | Replaced hardware keeps the device's name, contacts, rules and history on the new unit — no re-import, no orphaned history. |
| “Does the key let you change our network?” | SmartTile issues no configuration writes to UniFi. Delete the key on the console and access ends. |