Problem description

GLPI network discovery and SNMP inventory currently rely on a fixed, generic set of OID definitions and parsing logic maintained in core.

While this works well for common switches and printers, it does not scale well to modern environments that contain many other network-connected, non-agent devices, such as:

• IP cameras

• Access control / biometric devices

• Wireless access points

• IoT / OT equipment

• Vendor-specific or rebranded hardware

As a result:

• Devices are often detected but poorly identified

• Vendor-specific SNMP data is not parsed

• Administrators cannot extend or customize discovery behavior

• Community knowledge about device SNMP structures cannot be reused

Proposed enhancement (scoped MVP)

Introduce extensible, vendor-based discovery profiles that allow GLPI to apply targeted SNMP queries after basic device identification, without changing the existing discovery workflow.

This proposal is intentionally limited to enhancing discovery data, not automatic asset creation.

Minimal viable design

1. Vendor identification hints (non-authoritative)

During discovery, GLPI may already collect:

• MAC address

• sysObjectID

• sysDescr

Use these existing values to derive a vendor hint (best-effort), for example:

• MAC OUI lookup

• sysObjectID prefix match

• Simple sysDescr pattern match

No single signal must be authoritative.

2. Discovery profiles (read-only definitions)

A Discovery Profile defines:

• One or more vendor/device identification hints

• A list of additional SNMP OIDs to query

• Optional parsing rules for returned values

Profiles:

• Are applied only if identification hints match

• Fall back safely to generic discovery if unsupported

• Do not change existing SNMP logic unless explicitly matched

3. Extensibility (initially admin-managed)

For initial scope, profiles could be:

• Stored as configuration files

• Loaded at discovery time

• Enabled/disabled by administrators

This avoids:

• UI complexity

• Security concerns

• Runtime plugin execution

Out of scope (explicitly)

To keep this proposal feasible, the following are not requested:

• Automatic asset creation or conversion

• Changes to the asset model

• Replacement of existing SNMP discovery logic

• Vendor-specific UI elements

• Complex fingerprinting heuristics

Benefits

• Improved identification of non-agent devices

• Richer SNMP data for supported vendors

• Reduced number of poorly classified unmanaged devices

• Community-contributed discovery knowledge becomes reusable

• Future-proof foundation for modern networks

Why this is safe

• Discovery remains non-destructive

• Profiles are opt-in and additive

• Generic discovery behavior remains unchanged

• No assumptions or guesses are required

Example use cases

• IP cameras exposing limited but useful SNMP data

• Wireless access points without a central controller

• Access control devices reporting status via vendor OIDs

• Network devices using non-standard MIB layouts

Closing notes

This proposal aims to extend GLPI’s discovery capabilities without increasing risk or complexity in the core workflow.

By introducing a small, controlled abstraction layer, GLPI can better adapt to modern network environments while remaining backward-compatible and conservative by default.

Problem description

GLPI currently allows converting unmanaged devices only into a limited set of core asset types (Computer, Network Equipment, Printer, Phone) via the Convert feature.

However, modern environments contain a rapidly growing number of business-owned, network-connected devices that:

• Cannot run agents

• Are not well represented by existing core asset types

• Still must be inventoried and managed

Examples include (non-exhaustive):

• IP cameras

• Biometric readers

• Door/access controllers

• Wireless access points

• IoT / OT devices

• Environmental sensors

• AV equipment

At present, administrators must:

• Manually create custom asset types

• Manually recreate assets discovered as unmanaged devices

• Manually correlate MAC/IP/ports after the fact

This results in:

• Duplicate effort

• Inconsistent workflows (Convert exists for some types but not others)

• Increased risk of data inconsistency

• Poor scalability in environments with large numbers of non-agent devices

Proposed solution

Introduce a generic, user-defined conversion framework that allows administrators to explicitly define how unmanaged device discovery data can be converted into custom asset types, while preserving GLPI’s existing data-integrity guarantees.

High-level concept

1. Administrator creates a custom asset type (e.g. Camera)

2. Administrator defines a conversion template that maps:

   • Unmanaged device fields → asset fields
     (e.g. MAC, IP, name, network port, location)

3. GLPI validates the mapping:

   • Required fields present

   • No ambiguous or invalid mappings

4. Only after validation, GLPI enables:

   • Convert → <Custom Asset Type>
     for unmanaged devices matching the mapping

Conversion remains:

• Explicit

• Opt-in

• Deterministic

• Auditable

No automatic or heuristic guessing is introduced.

Why this is safe

This approach maintains the same (or stronger) safeguards as the existing Convert feature:

• No implicit field guessing

• No automatic promotion of discovery data

• Explicit administrator approval via mappings

• Required-field enforcement before conversion

• Clear lineage from discovery → inventory

In many cases, this would be safer than today’s manual recreation, which occurs outside GLPI’s visibility.

Use cases

• IP cameras discovered via network scans

• Access control and biometric devices

• Wireless access points not managed by a controller

• IoT / OT devices in enterprise environments

• Any business-owned network-connected device that cannot run an agent

Benefits

• Reduces manual inventory effort

• Improves CMDB accuracy and consistency

• Scales to modern network environments

• Avoids misuse of existing core asset types

• Aligns GLPI with current enterprise CMDB practices

Additional notes

This proposal is intentionally generic and extensible, not specific to any device category.
It replaces hard-coded conversion logic with a controlled framework while remaining fully backward-compatible with existing Convert functionality.

In ITSM / Release Management workflows, it is necessary to:

• Create a Release from a form

• Standardize release creation

• Avoid manual creation in Change / Project modules

The glpi-agent does not support notifying relevant personnel through webhooks only when CPU, memory, or hard disk changes

When adding the field Items->GLPI Object->Asset Computer it would be nice to be able to filter out inactive devices so the user can only pick from active devices

To search for devices that are nearing end of life we would like to be able to search asset fields for dates in the future, not just the past, tickets already have both positive and negative relative dates it would be nice if assets did too

There is a possibility to add feature in a future a button to list all the current actives in the logged user in helpdesk

Even its possible to send a notification when a active is assigned to an user.

We use the plugin and it works very well for us, but we are missing some functionalities. Could you consider including them in future versions?

• Improvements for the filter: make it search in the titles of projects, tasks, and subtasks.
• Export format: try to add some format like PDF or Excel for printing.
• Add the option to Show/Hide linked tickets in subtasks and make them clickable. The idea is that it should be optional. They should only be displayed if "Show" is selected.
• Add the possibility to add tickets directly, just like it currently allows adding new subtasks. Provide the option to add tickets in subtasks and have them linked to the project and the project task.

I'm using GLPI 11.0.4 as a Docker container.

In planning view I was not able to use the full available window height.

The only method I found was to manually edit line 93 of file planning.js inside the container, from 

height: options.height,

to

height: '100%',

and fully refresh the page in browser.

This value seems to have a default value set at line 55

height: GLPIPlanning.getHeight,

This behaviour is a bit annoying because I have to reproduce this modification at every containter recreation.

It should be nice if that value was parameterized somehow as many other parameters are already.

I’m trying to implement sequential approvals on Enterprise Service Management forms/tickets. I attempted to use Business Rules, but as soon as I add an additional criterion—such as Description → Contains, alongside Approval → Granted—the rule fails to trigger.

My requirement is to have different approvers for different types of service requests (e.g., recruitment requests, termination requests, leave requests). At any stage, if an approval is refused, the approval flow should stop and the ticket should close.

The challenge is that I cannot add additional criteria in the Business Rule to differentiate request types; as a result, the rule would apply globally to all tickets where approval is granted, forcing me to configure only one user or one group as the second-level approver—which defeats the purpose. ESM forms require different approvers depending on the request type.

Has anyone managed to solve this using Business Rules?

My other option is to explore configuring Entities, which in theory should work—but so should the rules, at least on paper. :(