Unify community and enterprise inventories with sensitive data obfuscation

[edospadoni]

Description

Currently NethServer sends 2 different inventories:

• Community: very complete, used by phone-home
• Enterprise: structured to satisfy the current my portal

For the new my portal, it would be useful to use the complete phone-home inventory, adding:

• Cluster ui_name
• Node ui_name, role (leader, worker), uptime, kernel, timezone, fqdn, public_ip (ipv4 e ipv6)
• Module ui_name, app certification level, domain references from Traefik routes, user_domains (array if multiple domains to same module)
• Domain details
• core and module updates available

Privacy concern

Some labels may contain sensitive data such as customer names, location names, or other identifying information.

Proposed solution

Use a single inventory with obfuscation of sensitive data for community (non-registered) installations.

To maintain data consistency (e.g., tracking that a domain has X connected elements), it is proposed to use a seed for random generation. This way, data remains consistent and traceable even when obfuscated.

Current example

Enterprise inventory (my):

{
    "node_id": 15,
    "domain_id": "ad.nethesis.it",
    "module_id": "webtop9",
    "instance_of": "webtop",
    "total_users": 103,
    "active_users": 82,
    "total_groups": 33
}

Community inventory (phone-home):

{
    "id": "webtop3",
    "version": "1.5.4",
    "name": "webtop",
    "node": "6"
}

There is information present in one but not the other (e.g., version, domain_id). The goal is to have all information in a single inventory, useful for both statistics and aggregate metrics.

Benefits

• Single data source easier to manage
• Ability to create aggregates also for community
• My portal can then decide which information to display
• Consistent data between the two modes

See also

• Discussion https://mattermost.nethesis.it/nethesis/pl/3a8zpom8y3b4fgsuiz5847g5xa

[DavidePrincipi]

In Testing

• core 3.17.0-dev.7
• traefik 4.1.3-dev.1

Test cases

With the Testing versions,

1. check the inventory contains the new fields
2. check data is pseudonymized without subscription
3. check the data format matches new MY portal expectations

[edospadoni]

LGTM