CMDB & Assets
Seredina's CMDB has three related but distinct screens: Assets (the inventory itself), Devices (the endpoint agent that feeds it automatically), and Equipment Catalog (the hardware models you can assign to an asset).
Assets
Each row is a piece of IT equipment, with its type, status, IP, hostname, serial number, manufacturer/model and operating system. The tabs above the list group the types:
| Group | Types |
|---|---|
| Computers | Workstation, laptop, server, tablet, mobile device |
| Network | Switch, router, firewall, Wi-Fi access point, other network device |
| Peripherals | Monitor, peripheral (keyboard, mouse, headset), docking station, printer, scanner, projector, IP phone, IP camera |
| Infrastructure | Storage (NAS/SAN), UPS |
An asset can be linked to the tickets that affect it (from the ticket itself, see Tickets). That builds up a history of what problems each piece of equipment had.
Inventory details
Besides the technical data, each asset can record:
- Asset tag: your own inventory number, such as TI-00042. Searchable.
- Assigned to: the person who uses it, one of your contacts. Their page lists the equipment they have.
- Location: floor, office or rack.
- Connected to: the equipment it plugs into, such as a monitor or dock to its computer, or a switch to its router. A computer's page lists what's connected to it, and + Add registers a new peripheral already connected.
- Purchase: supplier (from the directory), date, cost and currency, and when the warranty ends. An expired warranty is shown in red.
- Notes.
Monitors, peripherals and docks don't show network or operating system fields.
How an asset reaches the inventory
- Manual — created by hand from New Asset.
- Agentless discovery — scanning a network range (TCP + SNMP) from the Assets screen itself. Finds whatever responds on the network, without installing anything on the target machine — with that technique's logical limitations: if a machine has its firewall closed or SNMP disabled, it won't show up.
- Agent — see Devices below. Brings more detail than agentless discovery because it runs inside the machine, not from the outside.
The network scan recognizes switches, routers, firewalls, access points, storage, UPS, printers, IP phones and cameras from what each device reports over SNMP. The agent files laptops and tablets as such. It also registers the monitors it sees on a computer, matched by serial number, as Monitor assets connected to it. When a monitor moves to another computer, it follows; nobody has to type serial numbers in.
Devices
Devices is where you generate the install command for the Seredina agent. It's a single binary for Windows, macOS and Linux (x86-64 and ARM64), and it runs as a service. Every hour it reports:
- Hardware and software inventory, and the OS version.
- Whether the disk is encrypted.
- Antivirus status.
- The devices it sees on its local network.
To add a device:
- Click Generate enrollment command.
- Pick the device's operating system.
- Run the command there as administrator. It's valid for 15 minutes, and works once.
| System | Where to run it |
|---|---|
| Windows | PowerShell, opened as administrator |
| macOS / Linux | A terminal (it uses sudo) |
| Already downloaded | The agent's folder, for a device without internet access |
The command downloads the agent, checks its SHA-256, enrolls the device, and starts the service. If your server uses a private certificate, the command also carries its CA. The agent then trusts only that CA, and never turns off certificate verification.
If your devices can't reach GitHub, copy a release's files (install.sh, install.ps1, the archives and SHA256SUMS) to an internal web server. Set AGENT_DOWNLOAD_URL in .env to that folder's address. The commands then download from there.
seredina-agent status shows whether a device is enrolled and whether the service is running. seredina-agent uninstall --purge removes the agent. To update the agent, run seredina-agent update as administrator: no token needed, and the device keeps its record. See Updating the agents.
What the agent reports
Every hour the agent sends a full inventory. On the asset's page it's split into tabs, and Download JSON saves all of it:
| Tab | What's in it |
|---|---|
| Overview | Manufacturer, model and serial number; OS and build; CPU; memory; last boot; logged-on user; domain; IPs. Warnings for a pending reboot, an unencrypted disk, no active antivirus, or pending security updates. |
| Hardware | BIOS/UEFI; each memory module (slot, size, type, speed, serial); graphics; monitors with their serial; batteries with their health and cycles; printers. |
| Storage | Physical disks (SSD, HDD or NVMe, bus, serial, health) and volumes (filesystem, free space, encryption). |
| Network | Each adapter with its MAC, IPs, gateway, DNS, DHCP and speed. |
| Software | Every installed program with its version, publisher and install date, with a search box. |
| Security | Disk encryption, firewall, Secure Boot, TPM, UAC/SELinux/Gatekeeper, antivirus products and their state, security agents running (EDR), local administrators. |
| Updates | Installed updates (Windows KBs, macOS updates) and, on Linux, pending ones. |
| Services, Ports | For servers: each service's state and start mode, and what listens on which port. |
| Roles and VMs | Windows Server roles and recognized server software (SQL Server, IIS, PostgreSQL, nginx, Docker...), and the VMs and containers the machine hosts (Hyper-V, Proxmox, libvirt, Docker). |
The manufacturer, model, serial number and operating system also fill the asset's own fields, so Assets search finds a machine by its serial.
Servers
The agent works the same on servers: Windows Server 2012 or later, and any Linux with systemd (Ubuntu, Debian, RHEL, Rocky, Alma, SUSE...). A server is filed as Server on its own:
- Windows: the Server editions.
- Linux: machines without a graphical session.
If you change an asset's type by hand, your choice sticks. It only changes again if the machine's role really does change (for example, it's reinstalled as a workstation).
To see the inventory without sending it (for a machine you're checking, or to attach to a ticket), run seredina-agent inventory as administrator.
Inventory-only, by permanent design
The agent never executes anything remotely — no scripts, no software deployment. This isn't a temporary limitation of this version: remote execution needs an audited, signed delivery path this unfunded open-source project has no way to responsibly maintain. Agent releases aren't code-signed yet, either: SmartScreen or Gatekeeper may warn if you open the binary by hand, but the install commands aren't affected. See the MVP tour for the full reasoning.
Every enrolled device also shows up under Assets, tagged "AGENT" as its discovery source — it's the same inventory, just with an extra column showing where the data came from. Revoking a device from this screen cuts off its future check-ins without deleting its already-saved history.
Contracts, warranties and licenses
CMDB → Contracts tracks support contracts, warranties, software licenses, leases and subscriptions. Each one has:
- the supplier, picked from the directory along with the person to call there, or typed in;
- the contract or order number;
- start and end dates;
- the cost (one-time, monthly or yearly, in any currency);
- seats, for licenses;
- the assets it covers.
An asset's page lists the contracts that cover it.
Payments and reminders
To be reminded of a contract's payments, set its Payments section:
- the frequency: monthly, every two months, quarterly, every six months, yearly or one-time;
- the next due date;
- the amount, if it isn't the contract's cost;
- how many days before to remind (5 by default);
- the addresses to remind, such as finance or IT.
Seredina emails those addresses a reminder when a payment enters that window. If the due date passes with no payment recorded, it emails them once more. Each reminder is in your company's language and look, set in Customer emails. People whose role manages assets also get a Contract payment due notification, in-app and by email if they turned it on.
When a payment is made, click Record payment. Enter the date, the amount and an invoice or transfer number. The contract moves to the next due date, keeping the day of the month: a contract paid on the 31st is due on February 28, then March 31. Each contract keeps its payment history, and a mistaken entry can be undone.
The page header counts payments due soon and overdue ones. The Payments due soon and Overdue payments filters list them.
Each contract shows whether it's active, ending soon, or expired. The page header totals what's ending soon, what has expired, and the recurring cost per year for each currency. Remind days before (30 by default) controls the renewal reminder: once a contract enters that window, everyone whose role can manage assets gets a Contract ending notification, in-app and by email if they turned it on in their notification preferences. There's one reminder per end date; changing the end date (a renewal) re-arms it.
Don't put license keys in these fields. They're visible to anyone who can see assets, and included in data exports.
Equipment Catalog
A catalog of manufacturers and hardware models (for example, "Dell" → "Latitude 5540") — separate from individual assets. Assigning a catalog model to an asset pre-fills its default type, but the asset keeps its own independently editable type field: the catalog suggests, it doesn't force.