Aurora 99 Farms — Device Fleet HealthAurora 99 Farms — Salud de Flota

H2100004308 SILENT 14d

No chart data in the 14-day window. Either decommissioned, never reported through public_api, or recently registered without telemetry. (Note: at this site the gateway proper appears not to push chart data — its paired sink does. Check the sink record for connection history.)Sin datos en 14d.

SINKH2100004308 (sink) ONLINE

First seen in window:Primer visto: 2026-05-16T23:01:00Z  ·  Last seen:Última: 2026-05-22T23:01:00Z (0.0 h ago)  ·  Days with data:Días con datos: 6/14  ·  Max gap:Hueco máx: 23.4 h

14-Day Connection TimelineLínea de Tiempo 14d

Long Disconnects (>1h)Desconexiones Largas

H2100004285 SILENT 14d

No chart data in the 14-day window. Either decommissioned, never reported through public_api, or recently registered without telemetry. (Note: at this site the gateway proper appears not to push chart data — its paired sink does. Check the sink record for connection history.)Sin datos en 14d.

SINKH2100004285 (sink) ONLINE

First seen in window:Primer visto: 2026-05-20T23:01:00Z  ·  Last seen:Última: 2026-05-22T22:54:00Z (0.1 h ago)  ·  Days with data:Días con datos: 2/14  ·  Max gap:Hueco máx: 23.0 h

14-Day Connection TimelineLínea de Tiempo 14d

Long Disconnects (>1h)Desconexiones Largas

H4110005101 (repeater) ONLINE

First seen in window:Primer visto: 2026-05-21T23:00:00Z  ·  Last seen:Última: 2026-05-22T23:01:00Z (0.0 h ago)  ·  Days with data:Días con datos: 1/14  ·  Max gap:Hueco máx: 17.6 h

14-Day Connection TimelineLínea de Tiempo 14d

Long Disconnects (>1h)Desconexiones Largas

H4110005078 (repeater) ONLINE

First seen in window:Primer visto: 2026-05-20T23:01:00Z  ·  Last seen:Última: 2026-05-22T22:57:00Z (0.1 h ago)  ·  Days with data:Días con datos: 2/14  ·  Max gap:Hueco máx: 22.9 h

14-Day Connection TimelineLínea de Tiempo 14d

Long Disconnects (>1h)Desconexiones Largas

H4110005097 (repeater) ONLINE

First seen in window:Primer visto: 2026-05-20T23:01:00Z  ·  Last seen:Última: 2026-05-22T23:01:00Z (0.0 h ago)  ·  Days with data:Días con datos: 2/14  ·  Max gap:Hueco máx: 22.9 h

14-Day Connection TimelineLínea de Tiempo 14d

Long Disconnects (>1h)Desconexiones Largas

Silent Legacy RepeatersRepetidores Legado Silenciosos

H1300004576, H1300004553, H1300004769, H1300004761

Likely older-generation Wirepas hardware that's been retired. Worth confirming whether they were physically removed or are still on-wall but not powered.Hardware Wirepas antiguo. Confirmar si removido o sin energía.

Headline: 20 of 20 active mesh radios pass — every reporting device sits above the -75 dBm OK floor (median p50 -54.0 dBm). On this dataset alone, the new-building radio-blockage hypothesis does NOT show up as a measurable RSSI degradation. But — the active mesh has only 20 reporting nodes and most are clustered in a handful of zones, so a hypothesis test (place 3-4 test radios at points along the new building wall and pull RSSI in 48h) is the right next step. Your call.RSSI fuerte en todos los radios activos. La hipótesis del nuevo edificio no aparece en los datos actuales pero merece prueba con radios de test.

Headline: All 17 battery-reporting Wirepas nodes are healthy (p05 ≥ 3.5V, well above the 3.4V OK threshold). Lowest single-device p05 is 3.54V. No imminent Li-SOCl₂ replacements due. Repeaters, atm14, sinks and gateways are line-powered and report null battery as expected.Todos los nodos con batería >3.5V. Sin reemplazos inmediatos.

Healthy fleet floor: median p95 travel time is 0.07 s. The full distribution: 15 OK, 1 warn, 3 critical. The three critical-travel devices (H3210019837, H3210019903, H3210019873) all live in F3 or the unassigned pool and show p95 in the 76–181s range. H3210019903 is the one to look at first — it has a 119.8s travel time AND a link-quality reading of 0, which is the signature of a node stuck in a retry loop.Mayoría sub-segundo. Tres nodos críticos en F3 (76–181s p95). H3210019903 primero — bucle de reintento.

Note: this is Wirepas-reported node → sink travel. The downstream sink → gateway → AROYA cloud leg isn't exposed per-device in the public API; the "last seen" timestamp is the closest proxy and sits at 0.0–0.1 h for all reporting devices right now.Viaje publicado es nodo→sink; sink→nube no expuesto.

Healthy nodes route in 0.04–0.08s p95 (single-hop). The three travel-critical nodes (51–181s) are the only fleet members showing multi-second routing — these are device-localised issues, not global mesh congestion. Worth pairing this with the new-building hypothesis: are H3210019837 / H3210019903 / H3210019873 physically located on the far side of the new wall?Nodos sanos en single-hop (0.04–0.08s). Los críticos son localizados, no congestión global.

Based on the device location data: the active mesh telemetry is concentrated in F2, F3, F4 and the unassigned pool. F3 carries the heaviest critical load (6 of 12 devices critical; the two slow-routing nodes both live in F3). F4 has only 2 devices and both are critical. Flower Room 1 / Flower Room 2 (legacy room IDs) carry only silent devices and look like they've been migrated.F2/F3/F4 cargan el tráfico. F3 con 6/12 críticos.

For your second gateway across the building, I'd target a position that gives line-of-sight to F3 first. Your second sink already came online May 20 — confirm where it physically sits and whether F3's two slow-routing nodes (H3210019837, H3210019903) can route to it instead of jumping the long way to SINKH2100004308.Segundo gateway: posicionar con visión a F3 primero. El segundo sink ya online desde 20 may.

Don't forget the Ethernet upgrade. The action item from Call 1 — running Ethernet to gateways — would remove the upstream Wi-Fi variable from the failure tree. If the 23-hour outage on May 16-17 was upstream (WAN/Wi-Fi flapping), Ethernet eliminates it. If it was downstream (mesh starvation), the second gateway helps. Doing both gives you two independent improvements that don't rely on each other.No olvidar Ethernet a los gateways. Elimina variable de Wi-Fi upstream.

1. Confirm the gateway serial you shared on Call 2. Was it H2100004308 (the one whose paired sink had the 23.4-hour outage), or a different unit not yet registered in /devices/? The audit assumes it's H2100004308.Confirmar el serial del gateway. ¿Era H2100004308?
2. Test-sensor placement plan. Where would you place 3-4 test radios to confirm whether the new building is blocking signal? My guess from the data: one in F3 close to the wall, one in F4, one on the far side of the new build, one as a baseline near the primary sink.Plan de sensores de test para confirmar bloqueo del nuevo edificio.
3. Ethernet to gateways — timeline? Even a target month helps me time the next round of pulls.Ethernet a gateways — ¿fecha estimada?
4. Second gateway location — confirm where H2100004285 sits. If it's already on the new-building side, the recently-observed mesh extension is happening exactly where you needed it.Confirmar ubicación física del segundo gateway H2100004285.
5. Room cleanup. 51 of 81 devices have no room assignment in AROYA. Most of these are clearly active hardware (reporting in the last hour). Want me to draft a room-mapping pass for your next portal session?Limpieza de salas. 51 de 81 dispositivos sin asignación.
6. Legacy rooms / repeaters. Confirm whether Flower Room 1, Flower Room 2 and the four H1300004xxx repeaters can be archived.Confirmar si salas y repetidores legado pueden archivarse.

• Inventory: GET /public_api/devices/?facility=3415 · 81 devices.Inventario: 81 dispositivos.
• Per-device telemetry: 14-day chart pull in 2-day chunks (567 calls).Telemetría: 567 llamadas.
• Gateway/sink/repeater connection history: 14-day chart pull in 1-day chunks for the 11 gateway/sink/repeater devices to give you the high-resolution timeline.Historial gateways: pull diario de 14d.
• Flag thresholds (AROYA defaults): Battery p05 <3.0V critical / <3.2V warn · RSSI p05 <-95 dBm critical / <-85 dBm warn · Link <40 critical / <60 warn · Travel p95 >60s critical / >30s warn · Gap >60min critical / >30min warn.Umbrales AROYA por defecto.
• Why p05/p95: averages hide intermittent failures. A node solid 95% of the time but dropping out once a day looks fine on a mean; p05/p95 catches the bad moments.p05/p95 captura los momentos malos.
• Not in scope: any agronomic data (VPD, dryback, yield, crop steering). Device-side audit only.Solo lado dispositivo. Sin datos agronómicos.