Device compatibility¶
Overview of which devices have ready-made profiles in the community hub, and what to watch out for with specific vendors. Profiles are continuously verified against real hardware — if your exact model is missing, a profile from the same MIB world is usually a good starting point (clone + adjust, see Edit profiles).
Network¶
| Vendor | Profiles | Note |
|---|---|---|
| Ubiquiti UniFi | Switch, Access Point, Dream Machine, Network Controller | separate MIB world per device type |
| Aruba / HPE Instant On | Instant On switch, separate from ProCurve/Aruba CX | Instant On only speaks IF-MIB + standard PoE, no vendor-specific MIB |
| HPE ProCurve | dedicated profile | ProCurve/Aruba CX ≠ Instant On — not interchangeable |
| Cisco Catalyst | dedicated profile (IOS MIB world) | not compatible with SG/CBS — see below |
| Cisco Small Business (SG/CBS) | dedicated profile | its own MIB world, separate from Catalyst |
| MikroTik | separate router profile + switch profile | RouterOS MIB |
| Fortinet FortiGate | firewall profile | |
| SonicWall (Gen7) | firewall profile, verified against SonicOS 7.3 | see quirks below |
| OPNsense | firewall profile (REST API) | not via SNMP or shell — see quirks below |
Why Catalyst and SG/CBS are separate profiles: classic Cisco Catalyst (IOS) devices and the Small Business series (SG/CBS) speak different MIB trees. A profile built for one produces wrong or missing values on the other — so there are deliberately two profiles instead of one shared one. The same applies to ProCurve/Aruba CX versus Instant On.
Storage / NAS¶
| Vendor | Profile |
|---|---|
| Synology | NAS profile (SNMP, disk/volume aggregation, per-bay SMART) |
| QNAP | NAS profile |
| UGREEN NASync | NAS profile through the Linux agent (UGOS Pro is Debian-based): SMART disk health + temperatures, RAID (mdadm), btrfs pool integrity, system temperature/fans, SMB service, NAS overview with disk-bay panel; opt-in Docker + UPS (NUT) |
Power¶
| Vendor | Profile |
|---|---|
| APC | UPS profile |
Virtualization / servers¶
| Platform | Profile | Note |
|---|---|---|
| VMware ESXi | host profile (vSphere API) | read-only account, no shell needed — see quirks below |
| Proxmox VE | host profile · "Proxmox VE (API)" | Monitor over API token |
| Hyper-V | Windows Hyper-V host profile | |
| Windows Server | generic agent profile | CPU/RAM/disk/services/eventlog |
| Linux Server | generic agent profile | CPU/RAM/disk/services, also works over SSH (passive) |
Printers¶
Vendor profiles for HP, Brother, Lexmark, Canon, Xerox, Kyocera, and Ricoh — toner level, paper jams, page counters via SNMP (standard printer MIB, mostly compatible across vendors).
Device quirks¶
ESXi — monitored over the vSphere API, not via shell¶
VMware ESXi is monitored through the vSphere API. An account with the Read-only role is enough — shell access is not needed, and ESXi only lets Administrator accounts into the shell anyway (even with "Shell access" set on the account, ESXi rejects a non-admin account at SSH login). Measured: CPU, memory, datastore usage and reachability, running/stopped VMs, network uplinks, hardware sensors, connection state, uptime, and pending reboots. All checks of a device share one fetch per cycle.
The profile comes from the community hub and reads its metrics through a single, editable script (visible in the scripts library) — for "a VM is down" or "the host restarted" pick the metric built for that, with the right threshold direction (e.g. "VMs not running" instead of "VMs running", "Uptime" with a lower rather than an upper bound).
OPNsense — REST API instead of SNMP or shell¶
OPNsense is monitored through its REST API, not via SNMP and not via the shell. An account without shell access is enough — it only needs the "Lobby: Dashboard" right (which every account already has) plus the rights for whichever status pages are used (system status, gateways, services, interfaces, IPsec/OpenVPN). Reason for the move away from SNMP: OPNsense doesn't expose the SNMP counters common on firewalls (e.g. PF state counters) through its SNMP plugin, and the shell requires an administrator account just like on ESXi.
Synology — per-bay SMART values¶
In addition to disk status, the Synology profile reads the SMART state per bay: health as reported by DSM (normal/warning/critical/failing — the same classification DSM's Storage Manager shows under "Status"), remaining life in percent, and the bad-sector count. Whatever a drive doesn't report (SSDs typically don't report a bad-sector count, HDDs don't report remaining life) stays text "not reported" and is never shown as a number. The device view shows these values per bay under the "Device" tab.
Counter OIDs without absolute thresholds¶
SNMP counters like interface errors or STP topology changes count up since the device's last reboot — they'll eventually hit any threshold you set, regardless of whether there's actually a problem right now. Hub profiles deliberately avoid setting absolute CRITICAL/WARNING thresholds on such checks; instead they're shown as a trend/sum on the graph (purely informational, no alerting). If you build a counter check yourself: same principle — absolute thresholds on a monotonically increasing counter reliably end up permanently CRITICAL.
auto_add — not every check is active immediately¶
Some profile checks are marked auto_add=false and are not automatically
added to the host when you apply a profile. This mostly affects
management-port/SSL/SSH checks that would be chronically red on hardened
devices (e.g. a management interface only reachable from the LAN), or checks
that report different values depending on the device generation and
shouldn't be enabled blanket (e.g. SonicWall's RAM reading varies by
generation). The check's description in the profile explains when to enable
it manually — via Edit profiles or directly
in the host's service list.
Device snapshot + visualization¶
Device profiles in the hub usually include a check that fills the host's detail view with a matching visualisation ("Device" tab: switch port map, firewall panel, NAS disk bay, UPS panel — depending on the device type). Without such a check the detail page is limited to the plain check list. The snapshot does not have to come via SNMP: any check whose configuration carries a snapshot_type (e.g. "NAS Overview" in the UGREEN profile via the agent) fills the same panel — the matching visual is set automatically on import, and next to an existing snapshot provider no SNMP check is added that nobody could run on an agent host. The panel always shows the latest run.
Sonos speakers are monitored via their local HTTP interface (port 1400) — reachability with model, firmware, connection type (cable/Wi-Fi/home-theatre satellite), optionally Wi-Fi signal and battery (Move/Roam). There is no SSH or SNMP; discovery recognises Sonos across VLAN boundaries too (ports 1400/7000, device certificate on 1443).
MIB knowledge¶
Vendor-specific OID knowledge (custom MIB snippets, the sensor picker) is part of the host documentation: MIB snippets.
Next¶
- Community hub — search, import, and update profiles
- Edit profiles — adjust profiles locally
- MIB snippets — look up vendor OIDs