> ## Documentation Index
> Fetch the complete documentation index at: https://docs.maia.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Impact stream
Impact stream surfaces downstream dependency information at the moment of a pipeline failure, so you can see which pipelines and datasets are affected directly in context, without needing to manually navigate to the [Lineage](/docs/guides/data-lineage) view.
***
## What is Impact stream
Impact stream is a focused, run-specific view of the pipelines and datasets affected by a single pipeline failure. It's generated automatically from a failed run, so you can see the downstream impact of a failure the moment you notice it, instead of reconstructing it yourself in the [Lineage](/docs/guides/data-lineage) view.
Impact stream is available for pipeline runs with a **Failed** status.
***
## Open the Impact stream
You can open the Impact stream for a failed pipeline run from [Pipeline run history](/docs/guides/pipeline-run-history). On the **Your activity** page, locate a run with a **Failed** status and click the target icon 
next to the status badge.
This opens the **Impact stream** page, scoped to that specific pipeline run.
***
## Read the Impact stream
The **Impact stream** page has two panels: a pipeline tree on the left, and an Impact stream canvas on the right.
### Pipeline tree
The pipeline tree shows the hierarchy of pipelines and transformations involved in the run, starting from the root pipeline. Each node displays the pipeline or transformation name, its environment, and a colored dot indicating its status:
| Color | Status | Description |
| ----- | ---------------- | ------------------------------------------------------------------------ |
| Green | **Successful** | The pipeline or transformation completed without error. |
| Red | **Failed** | The pipeline or transformation encountered an error and didn't complete. |
| Gray | **Not executed** | The pipeline or transformation didn't run as part of this execution. |
Click the target icon next to a pipeline to expand it and view its child pipelines and transformations. The **Baseline execution** link at the top of the panel opens the original pipeline run that the Impact stream was generated from.
### Impact stream canvas
The Impact stream canvas visualizes how datasets and pipelines connect to one another, with arrows showing the direction of data flow. Each node is outlined in a color indicating its status:
| Color | Status | Description |
| ------ | ------------------------ | ------------------------------------------------------------------------------------------------------- |
| Green | **Successful** | The pipeline or dataset was processed without error. |
| Red | **Failed** | The pipeline or dataset encountered an error. |
| Gray | **Not executed** | The pipeline or dataset didn't run as part of this execution. |
| Yellow | **Potentially affected** | The pipeline or dataset is downstream of the failure and may be affected, but wasn't directly executed. |
| White | **Unaffected** | The pipeline or dataset has no dependency relationship with the failure. |
The pipeline tree uses three statuses (**Successful**, **Failed**, **Not executed**) because it only shows nodes that were part of the run. The Impact stream canvas adds **Potentially affected** and **Unaffected** because it also shows downstream datasets and pipelines that weren't executed but sit within—or outside—the failure's dependency path.
Each pipeline node also displays an **Impacted assets** count, showing how many downstream datasets and pipelines it affects. Dataset nodes display the source connector, dataset name, and column count.
***
## Impact stream and Lineage
Impact stream and [Lineage](/docs/guides/data-lineage) both show dependency relationships between pipelines and datasets, but they serve different purposes:
* **Impact stream** is scoped to a single failed run. It's generated automatically and answers "what does this failure affect, right now?"
* **Lineage** is a standing, comprehensive map of dependencies across all your pipelines and datasets. You browse it manually, at any time, independent of any specific run.
Use Impact stream to triage a failure as soon as you're alerted to it. If you need to investigate dependencies beyond the immediate downstream impact, open the full [Lineage](/docs/guides/data-lineage) view.