> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chargerdojo.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Verdicts, counts and run states

> The five verdicts, conformance and journey counters, run states, and how warnings and ChargerDojo recommendations differ from failed requirements.

Read the run state before interpreting its counts. A queued or active run has no final result. A report's verdict describes only its recorded scope; [coverage](https://docs.chargerdojo.com/guide/coverage.md) explains what lies outside it.

## The five verdicts

| Verdict | What happened | Next action |
| - | - | - |
| **All checks passed** | Graded checks passed with no failures, warnings or missing intended coverage. | Confirm the selected scope covers your release. |
| **Passed with warnings** | Graded checks passed; recommendations or observations produced warnings. | Read the warnings. They do not fail the run. |
| **Not compliant** | At least one graded required rule failed. | Investigate the first failed exchange, then retest. |
| **Partial coverage** | Nothing failed, but some intended evidence was ungraded or scope was withheld. | Read the named limits. The tested subset proves nothing about what was omitted. |
| **No checks ran** | No check received a conformance grade. This can include a run containing only ungraded checks. | Resolve the missing evidence or select runnable checks. This is not a pass. |

A failure takes priority over the other verdicts. With no failures, zero graded evidence takes priority over partial coverage, warnings and a clean pass. The app derives these labels from counts and evidence. The API returns summary fields and run status rather than a verdict word.

## Conformance counters

`data.summary` contains:

| Field | Meaning |
| - | - |
| `total` | All checks counted in this summary. |
| `passed` | Checks that passed their required validations. |
| `failed` | Checks that failed a required validation. |
| `warned` | Checks with a declined OCPI recommendation or an advisory observation. This count can overlap graded checks. |
| `unobservable` | **Not tested**. Nothing supplied a grade, or an earlier step left this check without what it needed. Read its reason. |
| `notApplicable` | **Not applicable**. Outside the endpoint's responsibilities, including an interface it does not advertise or a role ChargerDojo did not register with. |

For a conformance run:

```text theme={null}
total = passed + failed + unobservable + notApplicable
```

`warned` overlaps graded checks, so adding it again double-counts them. Not tested and not applicable are separate counts; neither is a failure. Older reports can omit `notApplicable`.

## Journey counters and authority

A charging journey counts exclusive outcomes:

```text theme={null}
total = passed + failed + warned + unobservable + notApplicable
```

Read the report's authority lanes as well as the aggregate counters. **OCPI conformance** checks protocol obligations. **Business checks** concern the chosen contract or configuration. **Observations** record behaviour that is not graded as an OCPI requirement. A partner agreement does not become an OCPI rule because it appears in the same report.

A journey's coverage and integrity fields also affect whether it provides a usable result. Follow the [journey guide](https://docs.chargerdojo.com/guide/run-a-charging-journey.md) for its completion policy; a suite's scoped-pass policy is not a substitute for a journey's required coverage.

## Run states

| `data.status` | What to do |
| - | - |
| `queued` | Wait for execution to start. |
| `running` | Follow progress; do not call the current counts final. |
| `recovering` | A recovery action is in progress. Read the current state again. |
| `recovery_required` | Read the server-offered recovery options in the journey. It still holds its run slot. |
| `completed` | Inspect summary, scope and integrity before accepting the result. |
| `failed` | Read the run error and recorded evidence. The state alone does not identify a failed protocol clause. |
| `cancelled` | Work was stopped; it is not a successful run. |
| `interrupted` | The run did not finish. A missing summary is not a summary of zeroes. |

The final states are `completed`, `failed`, `cancelled` and `interrupted`. An HTTP `200` from a report read can still contain an active state. Poll `data.status` with a deadline; use the [suite CI recipe](https://docs.chargerdojo.com/guide/run-a-suite.md#gate-your-build-on-it) rather than accepting a successful HTTP response alone.

## Why a warning is not a failure

OCPI requirement keywords distinguish required behaviour from recommendations. Declining a `SHOULD` is not a broken requirement, and a warning does not increment `failed`. A few warnings come from ChargerDojo's own rules rather than a `SHOULD`, and the summary names them as ours.

Sources: [OCPI 2.2.1 requirement keywords](https://github.com/ocpi/ocpi/blob/ca8c04c5c8cc10ad1be2a82f61b19377d98e2a67/terminology.asciidoc#terminology_terminology_and_definitions) and [RFC 2119 section 3](https://www.rfc-editor.org/rfc/rfc2119#section-3).

## ChargerDojo recommendations are not OCPI rules

A check can carry a **ChargerDojo recommendation**: advice from us where OCPI is silent, shown in blue with a bulb beside the status and explained when you open the check. It names its source and the date we read it. It is not a warning, never passes or fails a check, and changes no count and no verdict. The API returns it as `recommendations` on each check, and the MCP checks tool labels it as ours.

Response time follows two separate rules:

* **OCPI sets no response time.** When one of ChargerDojo's own limits stops a request (30 seconds of silence, or 60 seconds for the whole response), the check is **Not tested**, with the time it waited and the label `ChargerDojo limit`. It is never failed and never retried, because OCPI tells a client not to queue and resend a failed request. A run that reaches its own 10-minute limit stops as `interrupted`, keeps every check it already graded and sends no more.
* **We recommend answering within 10 seconds.** Real counterparties give up sooner: Plugsurfing's [Roam OCPI FAQ](https://developer.plugsurfing.com/docs/roam-ocpi-faq) documents a real-time authorization timeout configurable from 1 to 10 seconds. A slower answer is flagged on its check.

## Use the counts in CI

Zero failures alone is insufficient for a CI pass. Confirm completion, a nonempty valid summary and graded evidence, then apply the policy for that run type. [Diagnose a report](https://docs.chargerdojo.com/guide/read-your-report.md) when a required rule fails.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.