What it is
Deep Discovery re-scans your monitored devices with vendor-specific SNMP queries. A normal poll only reads standard metrics (traffic, status); Deep Discovery walks the vendor’s own MIBs to pull extended data — hardware details, model-specific tables — and stores it on each device’s profile.
In Observatory it runs as one bulk job over every eligible target. A target is eligible when it has SNMP enabled with a community string, a recognized vendor, and that vendor has Deep Discovery OIDs defined (built-in or from MIBs you uploaded).
The scan itself is executed by your Primary Local Agent, inside your network — the cloud only dispatches the task list.
How to run it
- Open Observatory. In the sidebar, next to the Devices header, click Deep Discovery.
- The button changes to Discovering…, you get a toast with how many targets were dispatched, and the button then shows live progress as
done/total. - When the job finishes (it is capped at roughly 2 minutes of polling), the button shows Done! and a “Deep discovery completed” toast appears.
- Open a device’s detail to see the refreshed vendor data.
You need edit permission on Observatory to trigger it; viewers can only watch progress.
When was the data taken
Deep Discovery data is a snapshot: it stays on the device profile until the next scan replaces it, even if the device later goes offline. Each scan is now dated — an access point’s detail shows “Deep data from …” in its header, so you can tell a fresh snapshot from an old one. Devices scanned before this feature existed have no date yet; the date appears after their next Deep Discovery run.
How it reaches the Agent: WebSocket vs REST mode
- WebSocket mode (normal): your Primary Agent is connected to the cloud, so the server sends it the task list directly. This works from anywhere — the Agent does not need to be on your machine.
- REST mode (fallback): if no Agent is connected via WebSocket, the server hands the task list back to your browser, which forwards it to an Agent running on your own machine (port 5050). This only works if the Agent is installed locally and your browser can reach it.
If neither path is available, nothing scans — the job will not progress.
How it differs from Wireless and UPS Deep Discovery
All three buttons use the same engine, but the scope differs:
- Observatory sidebar button: scans all eligible SNMP targets of any type, in one job.
- Wireless Monitor button: scans only access points and wireless controllers, and shows a progress dialog.
- UPS Monitor button: scans only UPS devices, also with a progress dialog.
Use the module-specific buttons when you only care about that fleet; use the Observatory one for a full sweep.
Troubleshooting
- “No eligible targets with SNMP + vendor OIDs” — your devices either have no SNMP community configured, no recognized vendor, or the vendor has no Deep Discovery OIDs. Fix the SNMP credentials and vendor on the device profile, or upload vendor MIBs in Auto-Provision.
- Job dispatched but progress stays at 0 — the Primary Agent is offline or never received the task; the job is then marked failed. Check the Agent in [[crearack—monitoring—fleet-manager]].
- Works in the office, not remotely — you are probably hitting REST mode (no Agent connected to the cloud). Install/connect an Agent so the WebSocket path is available; see [[crearack—terminal—local-agent]].
- No “Deep data from” date on a device — that device has not been deep-scanned since the date stamp was introduced; run Deep Discovery again and the date will appear.
- Per-vendor scan issues — see [[crearack—wireless—deep-discovery-troubleshooting]] for MIB-level troubleshooting that applies to all device types.
Related
- [[crearack—monitoring—que-es-observatory]] — the Observatory module
- [[crearack—monitoring—fleet-manager]] — which Agent runs the scan
- [[crearack—wireless—deep-discovery-troubleshooting]] — Deep Discovery for access points
- [[crearack—ups—groups-reports-troubleshooting]] — UPS module, including its Deep Discovery
- [[crearack—terminal—local-agent]] — installing the Local Agent
Véase también
- [[crearack—monitoring—que-es-observatory]]
- [[crearack—monitoring—fleet-manager]]
- [[crearack—wireless—deep-discovery-troubleshooting]]
- [[crearack—ups—groups-reports-troubleshooting]]
- [[crearack—terminal—local-agent]]