---
title: 'Fleet Inventory'
description: Inventory of connected corrector types and the archives they actually return, with discrepancies between expected and returned data highlighted.
section: AI Analytics
weight: 8
related:
  - ai-analytics/fleet-reports/passport-audit
---

import Alert from '@/components/docs/Alert.astro';

An inventory of the metering device types (correctors) actually connected and the archives they really return. The audience is operations and the support service: the report shows which equipment types are present in the fleet and whether the declared set of data matches what each device actually returns.

## What it shows

The main element is the **device type × archive kind** matrix. At the intersection of a row (type) and a column (archive kind: hourly, monthly, link sessions, events, settings changes, and so on) there is a color marker.

| Color | What it means |
|---|---|
| Green | data was expected and the device indeed returns it — everything matched |
| Yellow | data was expected but the device does not return it — a discrepancy worth investigating |
| Blue | data was not expected but the device returns it — something extra found, possibly useful |
| Gray | no data either by expectation or in fact |

Besides the matrix, for each device group the passport of the first polled node is shown: serial number, firmware version, and the upper and lower flow limits.

## How it's calculated

The report takes all fleet nodes and groups them by device type. For each type it polls several nodes (5 by default, for reliability — not just one) to find out the real list of available archives. If there are fewer than five nodes of that type, all of them are polled.

The obtained list is checked against a reference map of expected data sources per type. Discrepancies between what is declared and what the device returns are highlighted by color in the matrix. The report runs without parameters — the whole fleet is polled.

## Run parameters

The report runs without parameters: it automatically goes over the whole fleet.

| Parameter | Default | Meaning |
|---|---|---|
| (no form fields) | — | the whole fleet is polled |

## How to read it

First look for yellow cells — these are places where an expected archive does not arrive: there may be a device configuration problem or an outdated entry in the reference map. Blue cells mean the device returns data not expected from it — sometimes a useful source worth connecting in other reports. Green and gray cells are normal.

<Alert type="note" title="Limitations">

- For each type only a sample of nodes (up to 5) is polled. Rare settings on individual devices may not fall into the sample.
- If the equipment vendor updates the set of archives, the reference map must be updated manually — otherwise false discrepancies appear.
- Passport data is taken from the first polled node of the group and may not match other nodes of the same type.

</Alert>

## Related reports

- [Passport Audit](/en/platform/v3/ai-analytics/fleet-reports/passport-audit) — there the completeness of device passports and overall fleet health are assessed.