---
title: "Interpret a stopped or incomplete investigation | Proofline Help"
description: "Check the investigation’s state in its environment before using the conclusion. An alert can remain open after the investigation ends."
url: https://proofline.sh/help/investigation-terminal-state/
---

[Help center](/help/)

# Interpret a stopped or incomplete investigation

Check the investigation’s state in its environment before using the conclusion. An alert can remain open after the investigation ends.

## Distinguish execution state from alert closure

-   **Running**: Proofline is still reading evidence, or waiting for an alert's metric data before it reads it. The row updates when Proofline records the result.
-   **Failed**: the investigation ended without recording a result.
-   **Cancelled**: the investigation was cancelled before it finished.
-   **Stopped after failures**: the investigation stopped after repeated failures.
-   **No run recorded**: Proofline holds neither a result nor a run for this investigation.
-   **Likely cause**: Proofline names one candidate change as the likely cause, supported by cited timeline entries. The conclusion is limited to the evidence Proofline read.
-   **Ranked leads**: the result ranks candidate changes. A lead is a change to check first, not an established cause.
-   **Incomplete**: an ordinary investigation names no candidate change in the evidence Proofline read. This does not mean that nothing changed. For a fix re-check, the fix remains unchecked.
-   **No counterexample**: a fix re-check found no counterexample to the confirmed problem's requirement in fresh evidence. This never means that the fix is safe.
-   **Counterexample**: a fix re-check still finds that requirement violated.

An investigation of a Cloud Monitoring alert, a CloudWatch alarm or an Azure Monitor metric alert usually waits a few minutes for the provider's metric data before it reads the alert's metric window; [Alert investigations in Slack](/help/environment-investigations/#alert-investigations-in-slack) explains why and which alerts do not wait. The list shows **Running** throughout. On the investigation's page, an hourglass replaces the spinner beside **Running** and in **Conclusion**, which says **Proofline is waiting for the provider to publish the metric for the minute the alert opened in, and continues at** a time in UTC. When the wait ends, the spinner returns while Proofline reads the rest of the evidence.

A **Cancelled** investigation or one that **Stopped after failures** keeps any leads it reached. Those leads come from a partial investigation; their presence does not mean that it finished or established a cause.

Open an investigation to read its conclusion and evidence.

An empty list means that no investigation has run in this environment. It does not mean that the environment is healthy.

## Related guides

-   [Find an environment's investigations](/help/environment-investigations/)
-   [Create an environment](/help/create-environment/)
