Devices¶
A device is a single physical box - a switch, router, firewall, server, PDU, anything you rack and cable. This page covers creating one and reading its detail page.
Add a device¶
- Open DCIM → Devices → Devices in the sidebar and click Add device.
- Give it a name (must be unique within the tenant).
- Pick a device type - the hardware model. The box searches as you type; Search opens the full picker, which matches on name, model and part number, and filters by manufacturer, platform, tag, and artwork. That last one is more useful than it sounds: a rack elevation only draws properly for a type that has images, so has a front image answers "which of these will render". Don't have the type yet? Create it first on the Device catalog page; it only takes a moment and you'll reuse it for every device of that model.
- Optionally set the site, location, room, role, platform, software version (see Software target), cluster, status, serial number, asset tag and warranty end (see Support & warranty coverage).
- Optionally fill the built-in extras - comments, airflow, cooling method and latitude / longitude. Which optional fields appear is controlled by your administrator (see Built-in fields below), so you may not see all of them.
- Optionally pick Application groups and Profiles - the device profiles it gets, in order. Both are written with the first save. Add device on a device profile's page opens this form with that profile already picked.
- Save.
Adding many devices at once¶
A [a-b] range in the name creates one device per number in a single
submit: DK-RDS-[1-10] makes DK-RDS-1 … DK-RDS-10, and a zero-padded start
bound pads every name to its width, so [0001-0100] makes 0001 … 0100. A
line under Name previews it - Creates 10 devices: DK-RDS-1 … DK-RDS-10 -
and the button reads Create 10 devices. The rules are the same as for
components: one range per name, counting
up, at most 128 devices.
Every other field - type, site, role, status, platform, rack, tags, custom fields, application groups and profiles - applies to all of them. Fields that belong to one device must stay blank: a rack position, a DIN rail offset, a side-mount offset, serial number, asset tag, a stack position, the primary, secondary and out-of-band IPs and PRTG polls. A rack without a position is fine; the devices land in it unplaced. Place them on the rack afterwards.
It is all or nothing. If any name is taken, nothing is created and the
error names the taken ones; if one device fails validation or a permission
check, none are created and the error names which device failed and why.
Each device created is in the change log like a single create, and gets its
type's components, application groups and profiles as a single create does.
Creating takes add on devices at the chosen site, plus the
profile rights when profiles
are picked. On a planning task, a range plans one create per name.
Over the API, post the device payload with the range in name to
POST /api/devices/bulk-create/, with optional application_group_ids and
profile_ids. It answers 201 {count, results: [{id, name}]}, or 400
with the first failing device's field errors (prefixed with its name), a
detail summary and failed: [{name, errors}]. A plain
POST /api/devices/ refuses a name with a range rather than storing its
brackets.
curl -X POST https://danbyte.example/api/devices/bulk-create/ \
-H "Authorization: Token $TOKEN" -H "Content-Type: application/json" \
-d '{"name": "DK-RDS-[1-10]", "device_type_id": "…", "site_id": "…",
"profile_ids": ["…"]}'
Built-in fields vs. custom fields
Danbyte ships a curated set of common attributes as built-in device fields - including comments, location, cluster, airflow
(the device's own value overrides its type's default; the resolved value is
served as effective_airflow and drives the 3D room's airflow cones),
cooling method (likewise an override of its type's, served as
effective_cooling_method, air when neither sets one - see
Cooling) and latitude / longitude.
Comments, location, and the
coordinates are on by default (coordinates put a device on the
site map); cluster, airflow and cooling
method are opt-in under Settings → Device fields. Anything beyond that stays a
custom field, in line with the
zero-pre-filled-data philosophy.
While PRTG monitoring is on, the form has a PRTG
select: Follow policies (the default) leaves it to the PRTG policies
and the connection's Register all, Monitor in PRTG puts the device in
PRTG's plan whatever they say, and Exclude from PRTG keeps it out of
every plan. The API field is prtg_sync (inherit, always, never). A
muted line under it says what happens next: Planned changes wait in the
Review queue or Applied automatically. PRTG polls (prtg_host) names
the address or DNS name PRTG should poll instead of the primary address.
Changing either plans the device for PRTG at once, and the save says
Planned for PRTG or Being added to PRTG (prtg_plan on the
response). See
Registering devices and VMs.
The form's Topology card section sets which lines this device's card shows on the topology Diagram. Inherit (the default) takes them from the map's saved view, the device role or All devices, and previews the role's or All devices list; Custom gives the device its own list, and Name only shows just its name. See Card lines.
Under it, Icon gives this device its own icon on floor plans (icon),
over its type's and its role's; blank uses theirs. See
Icons when zoomed out.
You can also add devices in bulk from a spreadsheet - see Import & export.
Picking a device¶
Anywhere Danbyte asks you to choose a device - assigning an IP, adding an interface, terminating a tunnel or L2VPN, attaching a MAC address, adding a virtual-chassis member, setting a VM's host - you get the same device picker.
- Type to search. The picker is a searchable box; start typing a name and it narrows as you go. This is all you need when you know roughly what the device is called.
- Advanced search. When a name isn't enough - you're staring at hundreds of devices and want "the access switches in Amsterdam that are active" - click the sliders button beside the box. A dialog opens where you can filter by tag, manufacturer, device type, role, status, site, location, and region, alongside free-text search. Results come back in a table (name, type, role, site, status); click a row to select it.
Filtering by a region includes every site in that region and its sub-regions - pick "Europe" and you'll see devices in Netherlands → Amsterdam too, without selecting each child. All filtering runs on the server and respects your tenant and permission scope, so the picker only ever shows devices you're allowed to see.
The device page¶
Open any device to see its detail page. A slim header shows the name, status, tags, and description; below it is a row of tabs. If the device is a member of a switch stack, a Stack badge in the header (name, position, master) links to its virtual chassis - membership is set in the Stack membership section of the device's edit form.
The tab you're on is part of the URL (?tab=components), and so is the
sub-tab inside the Components tab (?sub=power). So
/devices/<id>?tab=components&sub=power links a colleague straight at a
device's power ports, and a reload, browser back/forward, or a trip through
another tab and back all keep your place. An unknown value in either param
falls back to the default tab instead of showing an empty pane.
The page draws in stages. Opened from the device list, the header and the
attribute cards appear at once from the list's row, while the device's own
read (GET /api/devices/<id>/) fills in the tab counts and the stack badge.
The heavier Overview cards - the panel, port utilization, support contracts,
the rack or cabinet drawing, the cooling path - each load on their own and
show a loading state in their own place until they do; a tab's content is
fetched only when the tab is opened. The header's compliance marker reads the
device's own status (GET /api/compliance/devices/<id>/, the
per-device compliance page's read), not the
tenant-wide evaluation, unless a list already fetched that.
Overview tab¶
The default tab lays the device's facts out in cards:
| Card | Shows |
|---|---|
| Device | name, status, role, platform (with the software version, where it came from and the platform's target) and its OS support bar, description, comments |
| Cooling path | shown once the device has a cooling intake or outflow: each intake's path from source through any CDU, each hop a link - see Cooling |
| Hardware | device type and its Hardware support bar, Support coverage, serial number, asset tag, warranty end, height (U), Size (W × H × D, see Size), airflow and cooling - each the device's own, else its type's marked (from type) - and for a device with antennas one Antennas line - count, type, gain, bands, connector - linking to the Hardware pane |
| Location | site, location, room, rack, position, face - or cabinet, rail and offset - and coordinates |
| Management | cluster, primary IP, its DNS name, and IP / interface counts |
| Support contracts | shown once a contract covers the device: each contract with its manufacturer, service level, end and state |
| Software | the first eight installed applications, each with its version and end-of-life lifetime bar, then View all N to the Software tab; No software recorded. when there is none |
| Profiles | the device's device profiles in order, each with its state pill, Re-apply and × to unassign, and Apply profile, which previews what a profile would write before it does |
| Sensors | shown once a PRTG sensor or channel is pinned to the device, under the monitoring strip: a pinned sensor as a tile with its state, primary value and the hour's sparkline, a pinned channel as a small gauge, lookup bar or value tile; a tile opens the sensor's panel, hover it to unpin |
Technical values (name, serial, asset tag, primary IP, DNS name) have a small copy button so you can grab them in one click.
The three support rows are lifetime bars: OS support from the platform's effective dates (entered, else its end-of-life cycle), Hardware support from the device type, and Support coverage from the device's contracts and warranty (Support & warranty coverage).
The Devices list has the same dates as an optional Next end of life
column (tick it in the Columns menu): the soonest of the three, or of an
installed application's end of life (Software tab), as the
lifetime bar, with which source it is, sorting by days left. The API carries
it as lifecycle_next: {source, end, start} (null with no dated end); a
coverage end has no start there. The whole estate, one row per source, is on
the Lifecycle overview.
Devices with ports also get a Port utilization card: a segmented bar plus
counts of connected (the port terminates a cable), reserved (its cable
carries the Planned status, or the uncabled port holds a direct
port reservation), and free ports, broken
down per kind (interfaces and front ports, plus virtual interfaces when they
are counted). A port can also be
marked connected (a one-click bolt action on the port rows, also a
checkbox on the interface / front-port / rear-port forms) when a cable is
physically in it but nobody has documented the cable yet - it counts as
connected, the row shows an Undocumented badge with a green tint, the
legend shows how many are undocumented, and the flag clears itself the
moment a real cable is attached to the port. The cable dialog lists such a
port as marked connected and lets you pick it, so documenting the cable
needs no unmarking first. Faceplates and photo panels draw
such a port dimmed by default; Settings → Admin → Faceplates → Light up
ports marked connected makes them draw it lit, like a cabled one. The same
card's Show interface prefixes on rendered faceplates prints the derived
prefix (Ethernet1/) in front of each port group; it is off by default
because on a dense switch it pushes the panel past its card. Label slots you
place yourself in the faceplate editor always show. Port labels prints
text inside each port's cage on the rendered faceplate, inside its marker on
the photo panel, and on the port quads in the 3D room: the port's own label
(the interface's Label field, A01, SS), the cable's label, the far-end
device's name, or the far-end port's label (a port's name is never printed
as a label) - and Label colour sets the text colour. The text is fitted to the marker and never leaves it, so a long
label only gets smaller; five characters or fewer stay readable on a dense
switch. On the rendered faceplate the port number stays, small, above the
label. A device's Port labels field (Inherit / Shown / Hidden) overrides
the deployment choice for that device - a patch panel wants them, a 48-port
access switch may not - and an interface's Hide label on faceplates
leaves that one marker blank, while its Label colour overrides the
deployment's colour for that port alone. On top of all that, the Labels
switch in the Panel header (and Port labels in the 3D room's View menu)
clears them off your screen only; it is remembered in the browser. Reserving works two ways: create
the cable ahead of time as Planned and the port counts as held, or reserve
the single port directly when the far end isn't known yet. Most useful on
patch panels and access switches, where "how full is this thing" is the
recurring question (GET /api/devices/<id>/port-utilization/).
What counts as a port. Physical interfaces and front ports, including management-only and disabled ones - they are real ports, free or not. Wireless radios are physical ports and count too. Left out:
- Virtual interfaces - SVIs, LAGs, loopbacks, tunnels and
sub-interfaces: anything ticked Virtual, or of type Virtual, Bridge or
LAG. The card says how many it left out (
12 virtual · not counted). Settings → Component details → Port counting → Count virtual interfaces counts them, deployment-wide. - Rear ports, always: a patch panel's rear is the back of the ports its front already counts, so a 24-port panel reads /24. The API still reports them, and the virtual interfaces, per kind.
- Interfaces whose status is Not present or Decommissioning leave the math entirely.
The same rule feeds every port figure: this card, the
stack card, the Port utilization page, the Devices
list Ports column, the cable picker's free-ports bar,
spec sheets and
port utilization rules. Each
response says which basis it used (count_virtual).
Changed in 0.17
Until 0.17 the total counted every interface and both sides of a patch panel: a 48-port switch with 100 SVIs read 38/148, and a 24-port panel /48. Totals now count physical interfaces and front ports, so most switches and panels read fuller than before. Interfaces of type Virtual or Bridge are now always flagged virtual, like LAGs, so they leave the faceplate and lose Connect and Reserve. The upgrade flags the existing ones, plus the SVIs, loopbacks and tunnels SNMP discovery created without a type, going by what their last poll reported.
The card is interactive: hovering a legend entry (connected / reserved /
free / undocumented) highlights the matching ports on the Panel above it -
on the photo faceplate and the rendered one alike - and clicking it
jumps to the Components tab with the port list pre-filtered
(?tab=components&sub=interfaces&cabled=free is linkable). The interface
and front/rear-port tables carry the same cabled-state chips
(All · Connected · Reserved · Undocumented · Free) for filtering by hand,
and Mark connected is bulk-editable, so ticking a whole undocumented
panel is one selection.
The estate-wide view lives at DCIM → Connections → Port utilization:
every device with counted ports, fullest first, with the same
connected/reserved/free split, its site and rack, site/rack/role/type
facets, search and export - so the patch panel about to run out is the
first row you see (GET /api/devices/port-utilization/). It lists only the
devices you can view. ?site=<id> or ?rack=<id> opens it with that facet
ticked. A device with nothing counted - a router with only loopbacks - has
no fill level, so it is not listed. The Devices list carries the same
number as a Ports bar column (like the prefix utilization bar), read for
the rows on screen only (?ids=<id>,<id>,…, up to 100 per request; a longer
page reads the whole roll-up), and
port utilization rules on the Alerts → Rules tab can notify when a
device's fill crosses a threshold - see
Monitoring.
Any custom fields you've defined for devices appear below the cards. When anything monitors the device - its own checks, Zabbix or PRTG - the monitoring card leads the tab: with problems it sits full width on top and lists the worst three, each a link to the check, the Zabbix section or the PRTG sensor's panel, then a bar per source; with none it is one line beside the first card (All 3 checks up · Zabbix up · read 45s ago). The header's monitoring pill is the overall state, with the sources on hover - see The monitoring bar. The Devices list also has a Monitoring column rolling that status up per device. Where the device physically sits - its rack elevation with this device highlighted - is drawn compactly (front and rear side by side) in the right column; a device in a DIN-rail cabinet gets the cabinet's plate instead, highlighted the same way. It's hidden for a device in neither. If the device's type has a rack-face image, that front/rear photo shows below the cards too.
Images¶
Uploaded photos and diagrams live on their own Images tab (rack shots, labels, cabling pictures, faceplate close-ups). Click Add image to upload; hover an image and click the trash icon to remove it - a confirmation asks first - or click an image to open the full-size original in a new tab.
Two layouts, chosen with the toggle beside Add image. The list
(default) names each file with its type, size, dimensions and when it changed,
sortable by name and searchable - and downloads no images at all, so a device
with fifty photos opens instantly; a row's link opens its original. The
grid shows the pictures, loading generated thumbnails rather than
originals. The choice is on the page's address (?images=grid), so a link
keeps it. The pencil on a row or card renames an image. Uploading and removing require change
permission on devices; everyone who can view the device sees the gallery
read-only. Files are stored under /media/ and served same-origin.
The same image gallery appears on rack, site, and location detail pages (on those it sits in the Overview) - one shared attachment system, each scoped to its object and gated by that object's change permission.
3D model¶
A device with a model - its own, else its type's - shows it on the Overview. The model is the 3D view of the Panel card, beside Photo / Rendered / Bare, and the card shows for a model alone too. Its header says whose model it is: From type or Own model. With no rack-face photo the model leads; with a photo the photo leads, and once you pick 3D it leads for you on every device until you pick a photo view again (kept in this browser).
The viewer turns with the mouse (drag to orbit, scroll to zoom). Its toolbar has Front, Rear and Top views, Fit (back to the three-quarter view), Ports to show or hide the mapped ports and Fullscreen. The ports are coloured by state as in the 3D room - live SNMP where the device was polled, else cabled by speed, free and disabled grey; hovering one names it, its far end and VLAN, and clicking an interface opens it.
The 3D model card manages the model; it says whether the device draws its own or its type's (From device type). Upload gives the device its own model, which overrides the type's for this device only; Remove takes it away again, so the device goes back to its type's. Uploading takes the same files - GLB, glTF, STEP, IGES, STL, OBJ - and follows the same rules as a device type's model; the Size, Up axis and Front settings show for the device's own model.
GET /api/devices/{id}/model/ downloads the model the room draws for the
device (its own, else its type's); POST and DELETE set and remove its
own, with change on the device. The device's model_3d is its own
model's URL, and model_3d_effective is what the room draws -
{url, source: "device" | "type", fit, up, front} - or null.
The card's Ports row counts the ports
mapped onto the model the device draws -
its type's (from device type) while it shows its type's model.
Map ports opens the same editor as the type's, listing this device's own
components; saving gives the device a mapping of its own, which replaces the
type's for this device, and Use type's ports goes back to the type's. A
device with its own model maps its own ports or none. GET / PUT
/api/devices/{id}/model-ports/ read and save it (its model_ports field;
null inherits). In the 3D room
the mapped ports show on the model in their state colour, and cables end at
them.
The rows shown in italics above - comments, airflow, cooling, location, coordinates, and cluster - are built-in fields whose visibility is admin-controlled; a device only shows the ones your administrator has enabled.
Size¶
A device can carry its own body size: Width, Height and Depth (mm) in the form's Hardware card, in tenths of a millimetre from 1 to 5000. Each one left blank takes its device type's size (the placeholder shows it), so set only what differs on this unit - a cabinet built to order, a desk, a shelf of gear modelled as one device.
The size in use - its own, else its type's - is what places the device:
- on a free floor plan its tile is its width × depth seen from above (grid plans keep one cell);
- in the 3D room a device with a width, height and depth and no 3D model stands as a box of that size in its role colour, and a 3D model fitted to size is fitted to it.
The Overview's Hardware card shows it as Size, W × H × D, marked
(from type) when the device has none of its own. The API takes and
returns width_mm, height_mm and depth_mm (null inherits) and serves
the size in use as effective_width_mm, effective_height_mm and
effective_depth_mm. Changes are in the device's change log, and a clone
keeps the size. DIN-rail placement in a cabinet still goes
by the type's size.
Local context data¶
Config context → Local context data on the device form holds JSON for this
device only, such as a default route or NAT settings two devices of a shared
role need. It is merged over every matching
config context, whatever
their weight, so its keys win; templates read it through config_context like
any other value. Leave it blank for none.
The API field is local_context_data: a JSON object, or null (an empty
object is stored as null; anything but an object is a 400). It is a JSON
column in device export and import, its changes are in the device's change
log, and a clone does not copy it.
Built-in fields¶
A handful of commonly-used attributes are promoted to built-in device fields rather than custom fields:
| Field | What it holds |
|---|---|
| Comments | Long-form notes about the device (multiline). |
| Location | The location within the site where the device lives, and the room in it. The Room field shows and hides with Location. |
| Cluster | The virtualization/compute cluster the device belongs to. |
| Airflow | Cooling direction - front-to-rear, rear-to-front, left-to-right, right-to-left, passive, or mixed. |
| Cooling method | How the device is cooled - air, liquid, hybrid or immersion. Blank inherits its type's; see Cooling. |
| Latitude / Longitude | Geographic coordinates of the device. |
Visibility is admin-controlled. Each field can be turned on or off
deployment-wide from Admin → Settings → Device fields (requires
users.manage), so the device form and Overview only show the fields your
administrator has enabled:
| Field | Shown by default |
|---|---|
comments |
Yes |
location |
Yes |
cluster |
No |
airflow |
No |
cooling_method |
No |
latitude |
No |
longitude |
No |
The setting is stored on the deployment singleton and read/written via
GET/PUT /api/deployment/device-fields/ (a flat object of those seven
booleans). A hidden cooling method still counts: a rack's cooling capability
is checked against the device's effective method whether the field shows or
not. Hidden fields disappear from the form and detail page, but any data
already set is preserved. If the setting can't be loaded, Danbyte falls back to
the same defaults.
Component tables¶
Every table on the Components tab - interfaces, console ports, console server ports, power ports, power outlets, cooling intakes and outflows, front ports, rear ports and inventory parts - has the Columns menu the list pages have: tick columns on or off, drag them into order, and Reset. The layout is saved per user for each table, and the menu also offers the list's other fields and custom fields as columns. Download exports the table as shown. The column headers sort, except on the inventory table, whose parts stay grouped under their parent.
The Add control for a sub-tab sits at the right of the sub-tab bar - one button when the tab adds one kind, an Add menu when it adds several. On Interfaces, Front ports and Rear ports the table's header stays in view while the rows scroll.
A cable trace opens a front, rear, console or power port on its own sub-tab
(?sub=front-ports, ?sub=rear-ports, ?sub=console, ?sub=power).
GET /api/devices/<id>/ counts the port tabs as front_port_count and
rear_port_count, and the Cooling sub-tab as cooling_count; since 0.18 hardware_count no longer includes front and
rear ports. Like the other tab counts they are 0 on the device list. On the
detail read every tab count comes from the one query that reads the device,
so the payload costs the same number of queries however much hangs off it;
the fields and their values are unchanged.
Other tabs¶
| Tab | What's there |
|---|---|
| IPs | Every address the device uses: assigned to it, to one of its interfaces, or set as its primary, OOB or secondary address while assigned elsewhere. Designation shows the slot; Assigned to shows This device, the interface (a link to Components) or the other object, with a hover note on an address that is here only through a slot. The tab count and the Overview's IP addresses count the same set (?device= on the IP list). |
| Components | Seven sub-tabs, each with its count: Interfaces (add, edit, and nest ports and attach IPs - see Interfaces), Console, Power, Cooling (intakes and outflows - see Cooling), Front ports and Rear ports (a patch panel's jacks and its back), and Hardware (device bays for child devices, module bays for line cards, antennas and serial-tracked inventory items). See Component tables. |
| Services | Application services running on the device. |
| Dependencies | What the device depends on and what depends on it, as trees, its peer links, and What if this fails? - see Device dependencies. Shown with view on device dependencies. |
| Software | Applications installed on the device, each at a version - add, edit and remove them here, or assign an application group. The tab shows the count (install_count on the device, list and detail). See Applications & installed software. |
| Flows | Application flows with an end at one of the device's addresses, or at a prefix or range that holds one. |
| Sensors | Only when PRTG monitoring is on, with the sensor count: a status strip, a card per PRTG sensor, and a panel with each sensor's channels as gauges and state bars. Before PRTG has sensors for the device it shows where it stands - excluded, not in PRTG, planned (with Add to PRTG now), needs attention, being added, or in PRTG without sensors - see Registering devices and VMs. The Monitoring tab gets one PRTG row linking here; its chevron opens whether the device is in PRTG, whether Danbyte made it, what registers it and what is planned, with Register in PRTG and Add to PRTG now - see Planned changes. |
| Contacts | People responsible for the device. |
| Config | Configuration context and rendered config, and the Oxidized backup when that is on. The config context lists its sources in merge order, the device's local context data last. |
| Journal | Free-form notes and a running log you write. |
| Change log | An automatic record of changes - who changed what, when. Also lists the changes to the device's dependencies, with an Object column. |
Status¶
A device's status (active, offline, staged, …) shows as a colored badge. The available statuses are yours to define - Danbyte doesn't ship a fixed list.
Custom fields¶
Need to track something Danbyte doesn't have a field for - an owner team, a maintenance window? Add a custom field for devices and it appears on every device's form and Overview. See Tags & custom fields.
Deleting a device¶
Use the Delete button in the device header. You'll be asked to confirm. Deleting a device removes its interfaces and any IP assignments on them.
To delete several, tick them on the device list and use Delete on the
selection bar: it first shows what would go and what is kept because it is
still in use (POST /api/devices/bulk-delete/).
When a device is linked to a PRTG device, both dialogs offer Also remove from PRTG - deleted in PRTG when Danbyte created it, otherwise paused unless you pick Delete it anyway. See PRTG: Deleting and retiring.
Front panel¶
The device Overview draws a front panel - the device rendered as hardware, at true physical scale: connector cages use their real millimetre dimensions (SFP narrower than QSFP, RJ45 taller than both) on an EIA-310-proportioned 1U bar. Ports lay out like the real faceplate (odd numbers on top, even below, banked in twelves), media groups split where the connector type changes, and color carries link state UniFi-style: sky for 10G+, emerald for 1G, amber below that, neutral for free ports, dashed for disabled. Trunk ports carry a top notch. Hover any port for its name, type, speed, far end (the device and port its cable reaches), VLAN (access/trunk + native), and IPs - the fields and their order are set under Settings → Component details - and click to open the interface.
Below the panel sits the Topology card: Paths lists one flat
end-to-end strip per cabled port (panels crossed front ⇄ rear, segments in
the cable's color); Map draws the 1-hop neighbourhood as the topology
Diagram does - each neighbour a card in its role's color with its card
lines, port names and addresses on their own cable, this device outlined,
LLDP links seen with no cable dashed (click one to make it a cable), and a
Legend chip in the corner (Trace maps);
Open in Topology opens the topology page
focused here.
Its header counts the runs and any LLDP links seen with no cable; a run
that dead-ends is marked Incomplete. On the
Interfaces tab, every cabled row carries a trace button (the same
strip in a dialog) and uncabled physical rows a ghosted connect button
that pre-seeds the cable form with the port as side A.
The same renderer draws every member of a
virtual chassis in the stack view.
The layout is automatic by default; when the device's type has a saved faceplate layout (built with the drag-and-drop builder), the panel follows that instead - including console, power, and aux ports placed on it. Layouts with a rear side add a Front / Rear toggle above the panel.
On a photo faceplate (a type with photo ports), the markers are work surfaces too, permissions allowing. An empty module bay takes a click to seat a module - the same install dialog as Components → Hardware, stamping the module type's interfaces onto the device - and the bay marker flips to occupied on save. The same works in the 3D room: the port card offers Install module on an empty bay and Edit part on a hardware marker (the same part editor the 2D faceplate opens for disk bays and PSUs). A free power / console / aux / front / rear marker connects a cable in place - see Cabling. Removing a module stays on the Hardware tab.
Part status. A hardware marker - a disk, a PSU, a fan - wears its part's status, and the status can be changed wherever the marker shows: on the device page's photo panel, on a rack's elevation, on a cabinet's plate, and on the part's card in the 3D room, a rack's 3D view and a cabinet's. The part's card lists the statuses the catalog offers inventory items as pills, the current one ticked; press one to set it. A right-click on the marker opens the same choices as a menu. The marker recolours at once in every view; if the server refuses, it goes back. The choices show only to users who may change inventory items, and the server checks each part against the user's scope, as for any edit; the change is in the part's change log. A port's state is not set this way - it comes from its cable and its enabled flag.
Racked devices also show a Rack card - the whole rack drawn with this device highlighted, linking to the rack page.
When the device has been polled over SNMP, each port also wears a small
live dot - emerald for oper-up, red for down, zinc for admin-down - and
the tooltip gains a live: line with the observed state and negotiated
speed. The overlay is read-only decoration from the monitoring collector:
observed facts are drawn over your intent, never written into it, so the
source of truth stays yours.
The panel's key¶
The speed ramp is always the full scale, <100M → 400G+, at a fixed width. It's a scale, and a scale only means something if it reads identically on every page - so it doesn't shrink to the speeds on the panel in front of you. (It briefly did. With two speeds present, two segments split a fixed-width bar into two enormous slabs, which looked like a different control rather than a shorter one.) It is the same scale the topology map's Speed colouring uses.
Changed in 0.17
The ramp gained a 100M tier. Everything below 1G used to share one amber tier, FE; 100M up to 1G keeps that amber, and anything slower - a 10M port, a 50M circuit - is now <100M, a darker amber. A speed stored as a bare number is read as kbps, as the server reads it.
The hardware key does adapt, because its entries are chips and a shorter list
is just a shorter list: a server whose photo panel is nothing but disk bays gets
Active · Empty, not the tenant's whole inventory-status catalog. Each status
is its pill, in the catalog's colour.
A virtual chassis draws one key for the whole stack, unioning what its members drew. In the 3D room the key is hidden entirely until something photo-anchored is in view.
Adding many components at once¶
Any component you add to a device takes a [a-b] range in its name and
creates one component per number - the same shorthand as a device type's
component templates. Type
Disk[1-5] and you get Disk1 … Disk5; a live line under the Name field shows
the count and the first/last name before you submit.
It works on every add dialog on the device page:
| Tab | Components |
|---|---|
| Interfaces | interfaces (Add interface) |
| Console | console ports, console server ports |
| Power | power ports (inlets), power outlets |
| Front ports | patch-panel front ports |
| Rear ports | patch-panel rear ports |
| Hardware | inventory parts |
Everything else on the dialog - type, speed, description, an outlet's inlet and
feed leg, tags - is applied to every component in the range, so a PDU's
Outlet[1-24] all hang off the same inlet in one submit. Ranges apply to
creating only: editing a component renames that one row.
Notes:
- The ports are created one at a time, in order. Names must be unique per device, so if one collides the error names the port that clashed and the ones created before it stay created.
- A leading zero on the first number sets the width:
Eth[01-04]creates Eth01 … Eth04.Eth[5-5]creates Eth5. - A range that counts down, spans more than 128 components, or appears twice in one name is refused - a name is never stored with its brackets. For more, reach for Bulk add on the Interfaces tab, which silently skips names the device already has. See Interfaces.
- Front ports advance a second field as they go. A front port claims its own
strand range on the rear port, and two of them may not share a strand - so the
range steps the Start strand along with the name.
Front[1-24]against a 24-strand rear port starting at strand 1 wires the whole trunk through in one submit; with a 2-strand connector each port takes the next pair. Pick the rear port and starting strand once and the rest follows. The API does the same for a rangedPOST /api/front-ports/and refuses the whole range if any port would overrun the rear port or overlap another. Splitter outputs all stay on position 1.
Component descriptions¶
Every component a device can carry - interfaces, console and console server ports, power ports and outlets, front and rear ports, aux ports, module and device bays, inventory parts - has a short free-text description for notes: what's on the far end, why a port is reserved, a ticket reference. It's a single line (255 characters). Fill it in on the component's add/edit dialog, read it back as a column in the component's table, and retype it across a whole selection from the bulk-edit bar.
A component template on the device type has one too, and it's copied onto every component stamped from it - so "reserved for out-of-band" written once on the type reaches every device built from it. Editing the concrete component's description afterwards doesn't touch the template.
Bulk editing components¶
Every component table - interfaces, console ports, power ports/outlets, front/rear ports on a device; VM interfaces; and the component templates on a device type - supports bulk editing. Tick rows and a floating bar appears:
- Edit opens a dialog where every field starts on Keep current; only the fields you explicitly set are applied to the selected rows. Booleans are tri-state (keep / yes / no), interface VLAN/VRF offer Clear, and tags can be added/removed without overwriting each row's other tags.
- Delete removes the selection after a confirmation.
Changes go through POST /api/<component>/bulk-update/ ({ids, fields}) and
bulk-delete/ ({ids}) - allow-listed fields per type, tenant-scoped,
audited in the change log like any other edit. Edit and Delete send more than
1000 rows 1000 at a time; Rename and Clone take at most 1000 rows (see
Large selections).
Spec sheet¶
Export → Spec sheet in the page header opens a printable PDF datasheet of the device - stat boxes, details, modules, interfaces with cable peers, comments and images - or, from the same menu, the hardware sheet: CPU, memory and storage totals first, then one table of parts per kind with the slot each sits in - or All in one, both on one sheet. The file is named after the device and its serial number. See Spec sheets.
PDF report¶
Export → PDF report… in the page header prints the device with the sections you pick - overview and custom fields, interfaces, IP addresses, cables, installed software, power and console ports, lifecycle, monitoring, journal and change log - each with its own columns. The device list's Import / Export menu and the bulk bar's Export menu make one report of the rows shown or selected: a table with the list's columns, a summary and, if you like, a page per device. See PDF reports.