Home/ User Guide/ Operator guide/ Administration

08 — Administration

The Amministrazione area is for administrators: user accounts and access control, reference data, alert and message configuration, translations, service credentials, and PyroWISE health. Most operators never need it; it's documented here for completeness and for the admins who run the cockpit.

Where these screens live. In the sidebar, the Amministrazione group only shows a couple of clickable links (Sign-up approvals and the Agentic entries). The rest are reached from the header toolbar — "Users & access" and "Settings" dropdowns (admin-only) — or by direct URL.

How permissions work here. Every action's permission is its route (e.g. /organisation/index). The admin role holds the super-permission (/*) and passes everything. A few screens additionally hard-check admin in code (portal visitor stats, the cockpit briefing editor, the agentic panels).


Users & access

Users — /user/index

Create and edit user accounts: name, contact, country (IT / SI), password reset, and — for admins — the user's RBAC roles (a multi-select). A professional mode toggle exposes per-user output preferences (always show regulatory references, attach bibliography to PDFs, include the scientific disclaimer, export raw calibration data). Reached from the header Users & access → Users.

The grid (updated September 2026) lists name, surname, e-mail, Status (active / inactive / deleted), Struttura / sede operativa — the operational agency the account is linked to, from the operational registry — the RBAC roles, Last login, the account expiry (red when lapsed, amber within 30 days) and the phone. Every header except RBAC role and Expire sorts by click, ascending then descending; Struttura and Last login sort with the accounts that have no agency / never logged in kept at the bottom in both directions, so a "who has not been in for a while" list is one click on Last login. Last login is the newest successful login recorded in the authentication audit log (user_auth_log), not a field on the account, so a user's first login after an account is created shows up on the next page load. The agency column accepts a text filter (funnel icon); admins also get a Simulate user action per row to sign in as that person for support, which is itself logged in the same audit trail.

Sign-up approvals — /user-approval/index

The moderation queue for self-service sign-ups — the one always-visible admin sidebar link. Tabs show Open (pending + info-requested) and each status. Open an applicant to see their profile and GDPR consents, then:

  • Approve & email user — activates the account, assigns roles (defaults to operator), sets an expiry (default +1 year) and emails a welcome note;
  • Ask for more information — sends a required question;
  • Reject — with a required reason.

A recent auth log is shown for context. See 01 — Getting started for the operator's side of this.

My account — /account/index

Any signed-in user's self-service page (not admin-only): profile, GDPR / marketing consents, a Manage API keys link, and a GDPR Cancel account action (soft-delete now, anonymise later).

RBAC — roles, permissions, assignments, routes

The access-control GUI, under the header Users & access dropdown. Five tools (module rbac):

ToolRouteWhat you do
Assignments/rbac/assignment/indexGrant/revoke roles & permissions to a user (find by e-mail, then a dual-list: Assign » / « Remove).
Roles/rbac/role/indexDefine roles as bundles of permissions/child-roles.
Permissions/rbac/permission/indexCurate permissions (in this app, mostly route strings like /incident/index, plus the super-permission /*).
Routes/rbac/route/indexAuto-discover controller/action routes and mint chosen ones as permissions (Refresh, then assign).
Rules/rbac/rule/indexBusiness rules (advanced; unlinked).

End-to-end: Routes (Refresh → assign to mint route permissions) → Permissions (curate) → Roles (build a role from permissions) → Assignments (grant the role to a user). Roles drive what each operator sees and which cockpit profile they get — see 01 — Getting started → Roles & profiles.


Reference data

Organisations — /organisation/index

Partner/member organisations (type, country, address, website, contact person). Users link to an organisation. Reached by direct URL / header.

Locations — /location/index

Named geographic sites with coordinates and terrain/forestry metadata. The location autocomplete and "get data" endpoints feed the alert matrix and the wildfire/incident forms, so keeping this clean matters downstream.

Don't confuse these admin Locations (reference geography) with the operational sites of agencies in the operational registry — different data, different purpose.


Alerts & messages

CMS pages (message templates) — /cms-page/index

Author the email subject/body and short message per alert type, using merge tokens ({NAME}, {SURNAME}, {NAMESURNAME}, {ORGANISATION}, {LOCATION}, {DATETIME}, {REASON}) that are filled in at send time. Reached from the header Settings → Cms Pages. These templates are consumed by the alert pipeline.

Alert matrix — /alert-matrix/index

The rules engine for notifications: which CMS-page/condition triggers an alert, at what frequency, for which locations, and who is notified — by user or by role — on which channel. A rule can also auto-open an incident. Each rule has Recipient rows (user · channel · active) and Roles rows (role · channel · active).

The alert matrix is the configuration; the live alerts operators read are at /alert/index — see 02 — Monitoring. Deliveries are recorded in the notification log.

Hazard-event alerts (TerraWise): severe natural-hazard events mirrored from TerraWise escalate through this same pipeline as alert type 26 — Hazard Event. Fan-out requires a CMS page for that alert type bound in an alert matrix rule with recipients (an alert_matrix row → cms_page with alert_type = 26) — the same onboarding as MeteoAlarm warnings. Without a rule the alert still appears at /alert/index; nobody is notified.

Daily briefing — /cockpit-briefing/index (admin)

Author the "Briefing of the day" that appears on the cockpit, per date and language (Markdown body). Briefings are normally generated by a nightly job; this screen lets an admin write or override one. Reached from the cockpit briefing widget's manage link.


Localization

Translations — /translation/index

The in-app i18n editor: edit translated UI strings per category, source message and language without a redeploy (an inline editable cell). This is where the help and UI strings get their IT / EN / SL / DE text.


Service credentials

API credentials — /api-credential/index

Personal API keys for the KF5 REST API and PyroWISE (sent as a Bearer token). Generate, Rotate (the old value stops working immediately) and Revoke keys. The full token is shown only once at create/rotate — copy it then; afterwards only a prefix is stored. Operators manage their own keys (from Account → Manage API keys); cross-user access is admin-only.

PyroWISE credential — /pyrowise-credential

The single shared X-PyroWISE-Internal-Key the cockpit uses to call the simulator (admin-only). The Save & validate action probes PyroWISE /healthz before storing — a bad key is rejected and the previous one kept. Re-validate re-checks the live key. A masked current value, source, last-validated status and recent history are shown. (Rotation runbook is in the README.)

API credentials (per-user bearer keys) and the PyroWISE credential (one shared internal key) are different things — don't confuse them.


Maintenance & legacy settings

  • Maintenance — /maintenance/index now redirects to the unified asset maintenance module (06 — Operations); the legacy grid is only reachable with ?legacy=1. (This is the maintenance log, not an application "maintenance mode".)
  • Technology / Technology type / Technology deployment — the legacy equipment registry, kept until the data is fully migrated into Field assets (06 — Operations). Technology type is under the header Settings menu; the equipment menu is hidden by default. Every save mirrors into the unified field_asset registry.

PyroWISE operations & health

Three read-only screens report on the PyroWISE service that powers simulation, benchmark and the per-run tools. They all show an amber "degraded" banner (never a 5xx) when PyroWISE is unreachable.

PyroWISE readiness — /pyrowise-readiness/index

The liveness and component readiness of the PyroWISE API. Reached from the header Settings → PyroWISE readiness. Shows a Liveness pill, a Readiness pill, the schema version, and a components table — each OK / Ready (green), Degraded (amber) or Unavailable / Not ready (red).

When something is red: PyroWISE (or that component) is down or misconfigured. Simulation, benchmark, smoke and run-explain tools will be degraded until it's green again. Check the PyroWISE service URL and the shared PyroWISE credential (above), and escalate if needed.

PyroWISE model status — /pyrowise-model-status/index

The active model configuration PyroWISE is running (direct URL): runtime version and git revision, FWI climatology mode, the fuel mappings in use, the calibrated priors (scope, season, version) and any event prior overrides. A few fixed operator notes flag known seasonal/gating caveats. Use it to confirm which calibration produced your runs.

Data inventory — /data-inventory/index

A read-only catalogue of every AOI layer descriptor, joined with PyroWISE's dataset catalogue (direct URL). Filter by AOI, group, status, view and readiness; the table lists each layer with its render type, status (ready green / preview amber / planned grey / error red), the AOI and applicable views, the build artifact state (building / stale / missing), dataset id, last-built time, hash, data-quality flags and any blocker. This is how you see which map layers/datasets exist per area and whether they're built. A panel at the bottom summarises EO mapping coverage (→ EO progression).

AOI activation previews — /aoi-profile-preview/index?aoi_profile_id=<id> (September 2026)

Every study area the engine publishes as not active can be previewed before it is switched on. That set is the engine's to change, not the cockpit's: on 2026-09-26 it held eleven profiles — Veneto, Istria (Croatia), Emilia-Romagna, Trentino-Alto Adige, Toscana, Umbria, Marche, Campania, Puglia, Lazio and Sicilia — while the five active ones (Karst Italy–Slovenia, Friuli Venezia Giulia, Slovenia, Sardinia and the Mediterranean catch-all) are what the switcher offers. Slovenia graduated to active in the meantime, so do not expect this list to stay still; the page itself always shows the current set. In the header's AOI switcher those profiles appear greyed out (hover for the reason); the eye icon beside the switcher opens the preview, and an in-page picker moves between the previewable profiles. The page shows the profile's identity and extent, the governed layer catalogue, the operational sites inside it, the activation blockers panel read live from the engine's activation-readiness endpoint (grouped by what is still missing on the engine side), and an Open governed 3D preview link (/map3-d?profile_preview=1) framed on the area. Previews are governed: nothing in them launches a run, sends an alert or writes a record, and opening one does not change the AOI your session works in. The feature is behind a host flag and its own permission — on some deployments (the Karst host as of 2026-09-26) the flag is off, and then the eye icon and the page are simply absent rather than empty; an administrator turns it on. Use it to judge whether an area's data plane is complete enough to ask for activation; the activation itself is an operator flip on the engine, tracked in docs/IMPLEMENTATION_STATUS.md §5.