# Audit Log

## Why audit logging matters

In AI and ML development, accountability and traceability aren't just best practices—they're essential.

With regulations like the EU AI Act, organizations must demonstrate control over their AI development and deployment pipelines. Audit logs provide the detailed, immutable record that lets you trace model origins and ensure development meets compliance requirements.

### What audit logs enable

* **Ensure compliance:** For regulated industries, having a verifiable history of all actions is mandatory. Audit logs act as an evidence trail, critical for compliance audits and legal investigations.
* **Debug issues:** When something goes wrong, you need to know exactly what actions were taken and by whom. Audit logs pinpoint the source of problems quickly and with clarity.
* **Maintain accountability:** In large teams, it's essential to know who is responsible for each change. Audit logs provide a clear picture of individual contributions and actions across your organization.

## Key capabilities

### Activity tracking

Captures all relevant activities with detailed information about who performed the action, what was done, and when it occurred.

Every log entry includes:

* Who performed the action
* What action was taken
* When it happened
* Where the action originated (IP address, location)

### Immutable records

Audit logs cannot be modified or deleted through the UI or API. This safeguards the integrity of the log and ensures you always have an accurate, unchangeable record of all actions.

### Long-term retention

Events are retained for 180 days from the date of occurrence, providing a comprehensive record over an extended period.

### Export capability

Export filtered logs in CSV format for integration with your reporting or audit processes.

### View limitation

The audit log view displays the 500 most recent entries. All entries are available via export.

## Accessing audit logs

Audit logs can be accessed in two ways:

### Organization view

Access through the organization dashboard under "Audit Log."

Shows all events across the organization with filtering by date range, project, or model.

![Audit log organization view](https://4109720758-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Ff3mjTRQNkASbnMbJqzJ2%2Fuploads%2Fgit-blob-ec9cafb5e4e46d8b78c703a656f94bdfa4e917b1%2Fauditlog.png?alt=media)

### Models view

Access through the "Audit Log" tab in model details.

Pre-filtered to show only events related to the specific model or model version. Filter by date range.

![Audit log models view](https://4109720758-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Ff3mjTRQNkASbnMbJqzJ2%2Fuploads%2Fgit-blob-02908be3b5cb059eb9e93498addcca5e8ca0d440%2Fmodelview.png?alt=media)

## Event information

Each audit log entry contains the following information:

| Field            | Description                                                  |
| ---------------- | ------------------------------------------------------------ |
| Timestamp        | The exact time when the event occurred                       |
| Event Category   | Area of Valohai where the event occurred                     |
| Event Type       | Specific action that was performed                           |
| Event Subject    | Resource affected by this event                              |
| Event Properties | Additional details specific to the event type                |
| Project          | Name of the associated project (if applicable)               |
| Actor            | Name of the user or system that performed the action         |
| Source           | Origin information of the event (e.g., IP address, location) |

> 💡 *Source information is only visible to organization administrators.*

## Tracked events

### Execution events

| Event Type | Properties                           |
| ---------- | ------------------------------------ |
| Run        | Commit ID, environment ID, step name |
| Stop       | -                                    |
| Delete     | -                                    |
| Rename     | Old name, New name                   |

### Task events

| Event Type | Properties                     |
| ---------- | ------------------------------ |
| Run        | Workload (execution, pipeline) |
| Stop       | -                              |
| Delete     | -                              |
| Rename     | Old name, New name             |

### Pipeline events

| Event Type | Properties         |
| ---------- | ------------------ |
| Run        | -                  |
| Stop       | -                  |
| Delete     | -                  |
| Rename     | Old name, New name |

### Datum events

| Event Type | Properties |
| ---------- | ---------- |
| Import     | -          |
| Download   | -          |
| Delete     | -          |
| Preview    | -          |

### Dataset events

| Event Type | Properties         |
| ---------- | ------------------ |
| Create     | Dataset name       |
| Delete     | -                  |
| Rename     | Old name, New name |

### Dataset version events

| Event Type | Properties |
| ---------- | ---------- |
| Create     | -          |
| Archive    | -          |

### Model events

| Event Type         | Properties                       |
| ------------------ | -------------------------------- |
| Create             | Model name                       |
| Archive            | -                                |
| Rename             | Old name, New name               |
| Description change | Old description, New description |

### Model version events

| Event Type | Properties         |
| ---------- | ------------------ |
| Create     | -                  |
| Approve    | -                  |
| Reject     | -                  |
| Tag change | Old tags, New tags |
| Download   | -                  |

### User events

| Event Type             | Properties           |
| ---------------------- | -------------------- |
| Login                  | -                    |
| Logout                 | -                    |
| Team membership change | Old teams, New teams |
| User name change       | Old name, New name   |

## Exporting audit logs

You can export audit logs for further analysis or record-keeping.

The exported CSV file includes all events within your selected date range, with one event per row.

## Built for AI governance

As AI governance evolves, so do the requirements placed on organizations developing and deploying AI.

Valohai's Audit Log provides a complete record of every action across your ML projects, making it easy to meet these demands.

Whether you're navigating compliance requirements like the EU AI Act, debugging production issues, or ensuring accountability within your team, the Audit Log delivers the transparency and control you need—right out of the box.