PRTG, placed in 3D

What you get

• Device layer: Ping, CPU and memory sensors colour each server in its rack; a red sensor shows as a fault on the device, its workloads and its tenant.
• Network layer: SNMP Traffic sensors give speed, utilisation and errors per port on the auto-laid-out topology. Fail a switch or link and see who loses every path.
• Power and Air layers: temperature channels from system-health sensors become rack-inlet colours; UPS sensors put load, battery and runtime (estimated when PRTG has no runtime channel) on the one-line diagram. Click the UPS → Fail → what goes dark; minutes to 27 / 32 / 35 °C are estimated.

What we read (today)

PRTG export field	DC Sim sheet.column	Layer
Device name (content=devices)	Devices.name	Device
Device tags such as rack:A01 and u:17, or the Location / Comments field (PRTG has no rack field)	Devices.rack, Devices.u	Floor
Group or probe name, if you group per row	Racks.row	Floor
Model text (a tag, the Comments field, or the type auto-discovery detected)	Devices.model, matched to the catalog	Device, Power, Air
Ping sensor status	Readings metric up	Device, Network
CPU Load sensor last value (SNMP, WMI or SSH)	Readings metric util_pct	Device
Memory sensor %	Readings metric mem_used_pct	Device
SNMP Traffic sensor channels Traffic In / Traffic Out	Readings item device:port, metric link_mbps	Network
SNMP Traffic sensor channels Errors In / Errors Out	Readings item device:port, metric port_errors	Network, Cabling
Temperature channels from system-health sensors (SNMP hardware, IPMI)	Readings metric inlet_c or temp_c	Air
Fan and power-supply channels	Readings metrics fan_rpm, psu_ok	Device
UPS sensors: load %, battery capacity %, remaining runtime	Readings metrics ups_load_pct, ups_battery_pct, ups_runtime_min	Power
Power-meter or PDU sensors with a watts channel	Readings metric power_w	Power, Token (tokens per kWh)

PRTG’s historicdata CSV has one row per timestamp, so it drops straight into the Readings sheet — time, item, metric, value — with the device name as item and the channel name as metric.

What we will read / write back (roadmap)

Item	Status
Read-only PRTG connector in the DC Sim agent: device tree and sensor values polled through the HTTP API with a read-only API token	Roadmap (today: CSV export)
Device tree → devices and rack positions automatically from tags	Roadmap
PRTG sensor states → DC Sim alarms with advice	Roadmap
Write back: rack and U tags on PRTG devices from the DC Sim design	Roadmap

Setup

1. Export from PRTG. With a read-only API token, call table.xml?content=devices&output=csvtable (columns: device, host, group, tags, location, comments) and table.xml?content=sensors&output=csvtable (device, sensor, status, last value). For a time series, download a sensor’s Historic Data tab as CSV or call historicdata.csv with the sensor id and date range.
2. Map the columns. Download the DC Sim Excel template and paste devices into Devices (rack and U from your tags), rack groups into Racks, sensor values into Readings. Headers match case-insensitively with synonyms (“Hostname”, “Rack name”, “U position”, “Model number”), so rename only what the importer cannot guess.
3. Import. Import → preview → report → apply. Up to 10 MB per file, 50,000 rows per sheet, values only; bad rows are listed and never block the import.
4. Read the check list — U overlaps, overweight racks, breakers above 80 %, unknown models — each with its 3D location; a re-upload shows what changed.
5. Ask to be told when the live, read-only PRTG connector is ready: da@mustardseedmt.com.

Nothing inbound

Today nothing connects to PRTG: you upload a file. The roadmap connector will run in the outbound-only DC Sim agent: one WebSocket out, no listening port, no notification webhook to expose, batches spooled to disk and replayed in order after an outage, and a PRTG API token with read-only access that never leaves your network.