Skip to content

Adding hosts

A host in Vesana is one concrete monitored machine or device — a Linux server, a switch, a UPS, a firewall.

Ways to create a host

Way When
Add Host (wizard or expert mode) A single device you already know
Discovery Scan the network, take over hits with tenant and device type pre-selected
Set up automatically Fresh instance: Vesana proposes devices and checks itself
Copy host A similar device already exists
API Several hosts by script — see API

This page describes "Add Host" and the "Add check" dialog. For the scan route see Discovery.

Wizard or expert mode

You open "Add Host" via the button top right on the hosts page, via the "+" in a tenant group of the device list, or from Discovery.

There are two views of the same dialog, switchable top right:

View For whom
Wizard (default) Three guided steps: Device → Checks → Connection
Expert mode All fields on one screen

Your choice is remembered per browser. Switching carries over all inputs — including individually selected checks; nothing is lost.

The wizard in three steps

Step 1 — Device

First the tenant (searchable dropdown), then "What do you want to monitor?":

Choice Meaning
Linux Server/workstation with agent — measures CPU, RAM, disk and services directly on the device
Windows Same, with agent
Something else Device without agent — switch, firewall, UPS, NAS, hypervisor … Checked from the server (active checker) or via a collector

Coming from Discovery, tenant and device type are already pre-selected.

Step 2 — Checks

Here you decide what gets monitored. There are two doors, shown as tabs above the selection:

  • Profile / template — device profiles that bring a set of checks (see Profiles & checks)
  • Individual checks — the catalog: ping, website, port, certificate, SNMP sensor, …

Both doors mix freely: profile checks and individual checks are created together; the summary line at the bottom shows how many profiles and checks are selected.

Profiles:

  • For Linux/Windows the recommended base profile sits in its own section at the top and is pre-selected — even if it has not been imported locally yet.
  • Several profiles can be combined, e.g. Windows base plus Hyper-V.
  • Each profile card shows the number of checks and which access the profile needs (SNMP, SSH, API). "Adjust checks" lets you deselect individual checks of a profile.
  • "Show all profiles" opens the full list including community suggestions from the hub. A filter controls the suggestions: Community: only new (already imported profiles do not appear twice), show all or hidden. Local profiles always stay visible; imported ones carry the "imported" mark.
  • A community profile that is not imported yet gets imported when the host is created — several of them one after the other.

For an agent device you can continue without a profile: the agent delivers base telemetry, checks can be added at any time.

Search: The search field applies to both doors at once (from v1.9.437). While typing, the tabs show the number of matches per side; if the current side stays empty, a hint names the matches on the other side and switches there via "Show" — if exactly one profile matches, it is selected directly. The search term survives the switch.

Step 3 — Connection

  • Name — display name, e.g. "Firewall HQ" or "srv-01". Below it, "Internal name:" shows the derived hostname live, with a pencil to adjust it and a live check ("taken" with suggestion / "available"). Details under Internal name.
  • Target — IP address or DNS name. If a selected check needs a target (e.g. Proxmox VE (API) or an SNMP check), the field is mandatory and appears in the "Still needed" list. If checks only target the network incidentally, you can create the host without a target — a hint tells you those checks will only run once a target is set.
  • Check via — appears only if at least one collector is assigned to the tenant: Active checker (the server checks itself) or Collector; with several collectors a dropdown.
  • Credentials — there are exactly three access types on a device: SNMP, SSH and API access. Their sections open automatically and are marked "Required for selected checks" as soon as a selected check needs them. Per section you choose "Stored credentials" (a credential stored in the tenant is applied on creation, no typing needed) or "Enter manually". API access consists of User and Password / secret — the same two fields take both an ESXi read-only account and a Proxmox token (token ID as user, secret as password). A check only asks for the access type it actually uses — a device API without a login (e.g. Sonos) doesn't ask for anything. Entered credentials are stored encrypted; the browser's password manager deliberately does not autofill these fields (it would otherwise insert a stale, unrelated password).
  • For Linux/Windows with agent there are no credentials: the agent connects to the server itself, token and install command follow after creation.

Bottom left, a "Still needed:" list shows live what is missing ("choose a tenant", "enter a name", "select at least one check", "Enter a target (IP or DNS) — the selected checks require it"). "Add Host" is never silently disabled — the list always tells you why a click does not do anything yet.

Description, site and site override are not part of the wizard — enter them in expert mode under "More details" or later on the host page.

Expert mode

All fields sit on one screen with a pinned footer. The flow matches the wizard, with these differences:

  • Device type as three pills Linux / Windows / Other. For Linux/Windows there is a checkbox "With agent (recommended)"; without agent or for "Other", "Check via" appears with Active checker and Collector.
  • Profiles are selected via checkbox; clicking the row only shows the details. An overview line "Selected:" lists all chosen profiles with their check counts — deselectable by click. Here too, several profiles and both doors (Profile / template, Individual checks) can be combined.
  • "More details" (collapsible): description (max. 500 characters), site (see Sites) and a custom site override with geo coordinates.
  • Monitoring modes (agent / collector / active checker) are additive: if you select a check that needs a mode not yet active, that mode is added automatically — an inline hint names it. That is how a Linux server with agent gets an additional SNMP check via a collector in a single pass.

Internal name

The internal name (hostname) is derived automatically from the display name as long as you do not touch it — after the first manual change the derivation stops. It is the name under which agent and collector track the host.

Renaming

When editing later: renaming the internal name can interrupt running assignments until agent or collector have loaded the new config. The display name, by contrast, can be changed safely at any time.

After creation

Success screen with agent token

For an agent device the success screen shows "Host "…" has been created." with token and install command (tabs for Linux/Windows) — see Agent installation. From here "Add more devices" goes straight into the next dialog, "Continue to host" to the host page.

If token generation fails, you see instead: "The agent token could not be generated — you can generate it at any time on the host page via "Setup agent"."

Skipped checks

If a selected check could not be created, the success screen lists "{n} checks were skipped:" with a reason per check:

  • needs agent mode — the host is not agent-managed
  • needs a target (IP or hostname) — not set on the host
  • was deleted in the meantime
  • already exists

You can add them any time via "Add check".

Duplicates

An internal name that is already taken is rejected by the server and shown as an error in the dialog. Together with the live check in the name field this prevents most duplicates before submitting.

"Add check" on the host page

On the host page you add checks via "Add check". The dialog uses the same two doors as creation — Profile / template and Individual checks — including the shared search across both (from v1.9.437).

Execution: The dialog always shows all three ways as tiles (from v1.9.437) — Agent, Collector, Active checker — each with its online state. Whatever the host does not have yet, you set up straight from the tile:

Tile If missing
Agent — checks directly on the device (CPU, RAM, services, logs) "Setup agent" opens token and install command; "Continue to checks" returns to the selection. An existing token stays valid unless you explicitly create a new one — that invalidates the old one
Collector — checks from inside the customer network (SNMP, ping, SSH) "Assign collector": with exactly one collector in the tenant it is pre-selected; with several you pick from a searchable list with online state. If there is none yet, a link leads to the tenant page
Active checker — the Vesana server checks by itself (ping, HTTP, SSL) "Enter target": IP address or DNS name of the device

Further fields in the dialog:

  • Target (optional) per check — default is the host's IP/DNS; enter a different address here (e.g. a second interface). If a check needs a target and the host has none, the input opens right away and is applied afterwards.
  • Check name — sits right in the row of the selected check and can be edited there. Pick the same type several times (say nine websites) and the checks get consecutive names automatically (“Website”, “Website 2”, …), also against checks the host already has. A custom name replaces the number.
  • If a selected check needs SNMP or SSH and the host has no credentials yet, the dialog asks for them on the spot.

To continue on the host page (editing checks, bulk actions, tags) see Host detail page.

Rules in the background

When checks are created, validations apply that you cannot bypass in the UI:

  • Agent checks → the host needs an agent token
  • Collector checks → the host must be assigned to a collector of the same tenant; a collector of another tenant is rejected (from v1.9.437). The active collector stays selectable for all tenants
  • Network checks (ping, port, SNMP, SSH …) → IP address or DNS name required
  • Interval between 10 s and 7 days
  • The check type is immutable after creation — for a different type, create the check anew

Copying a host

Instead of creating from scratch, right-click in the device list or structure column and choose Copy Host. The dialog asks for host name, display name, IP address and target tenant (default: same tenant). This is also the way to bring a device into another tenant — moving via the structure only works within its own tenant (see Hosts page & structure).

Next