Audit trail on Dataiku Cloud

On Dataiku Cloud, audit logs will be automatically served through connections installed on your DSS, so you can leverage them directly in your instance.

Access API node queries

An Amazon S3 connection named “customer-audit-log” hosting your API Nodes logs queries is automatically added to your instance when you activate the API nodes. No further action is required.

To import your logs as a dataset in your Dataiku’s instance:

  1. Add a Dataset > Cloud storages & Social > Amazon S3

  2. Select the corresponding S3 connection and the path in bucket

../../_images/connection_selection.png

  1. In the “Format / Preview” tab, select “One record per line” as Type and “utf8” as Charset

../../_images/format_selection.png

  1. In the “Partitioning” tab, activate the partitioning: add a time dimension corresponding to the period you want to partition on, in the example below we partition per day

../../_images/partitioning_api_queries.png

  1. Create the Dataset and access it in the Flow

Access instances’ & Launchpad audit logs

This feature is only available for customers.

Activate the extension “Audit Logs” in your Launchpad to add two Amazon S3 connections named: * “dku-audit-log” hosting your audit trail accessible from your Dataiku instance. * “launchpad-data” hosting Launchpad audit trail

By default only space-administrators can access these connections, you can edit this behavior in the connection settings.

Instances audit logs

Once the connection is available, to import the logs as a dataset in your Dataiku’s instance:

  1. Add a Dataset > Cloud storages & Social > Amazon S3

  2. Select the “dku-audit-log” S3 connection and the path in bucket

../../_images/select_auditlogs.png

  1. In the “Format / Preview” tab, select “One record per line” as Type and “utf8” as Charset

../../_images/format_selection.png

5. In the “Partitioning” tab, activate the partitioning: add a discrete dimension called “node_type”, and add a time dimension corresponding to the period you want to partition on, in the example below we partition per day

../../_images/partitioning_settings.png

  1. Create the Dataset and access it in the Flow

Launchpad audit logs

  1. Add a Dataset > Cloud storages & Social > Amazon S3

  2. Select the launchpad-data S3 connection and browse to set the “Path in bucket” to /audit-trail/

  3. If you want to partition the dataset, in the “Partitioning” tab, activate the partitioning and add a time dimension corresponding to the period you want to partition on. Fill in the Pattern /%Y/%M/%D/.* for a daily partition. The dataset can also be hourly partitioned with the following pattern /%Y/%M/%D/%H.*

  4. Create the Dataset and access it in the Flow.