Glue Data Catalog
Last updated
Last updated
This guide describes how to use the Glue Data Catalog data source in order to monitor a new ML Model in production.
We will assume that your model inputs, outputs and optionally delayed actuals can be queried exist as tables in Glue Data Catalog. This data source may also be used to connect to your model's training/test set to be used as a baseline for model monitoring.
Log into your AWS Console and go to the IAM console.
Click the Roles tab in the sidebar.
Click Create role.
In Select type of trusted entity, click the Web Identity tile.
Under Identity Provider, click on Create New.
Under Provider Type, click the OpenID Connect tile.
In the Provider URL field, enter the Aporia cluster OIDC URL.
In the Audience field, enter "sts.amazonaws.com".
Click the Add provider button.
Close the new tab
Refresh the Identity Provider list.
Select the newly created identity provider.
In the Audience field, select “sts.amazonaws.com”.
Click the Next button.
Click the Next button.
In the Role name field, enter a role name.
In the list of roles, click the role you created.
Add an inline policy.
On the Permissions tab, click Add permissions then click Create inline policy.
In the policy editor, click the JSON tab.
Copy the following access policy, and make sure to fill your correct region, account ID and restrict access to specific databases and tables if necessary.
Click Review Policy.
In the Name field, enter a policy name.
Click Create policy.
If you use Service Control Policies to deny certain actions at the AWS account level, ensure that sts:AssumeRoleWithWebIdentity is allowlisted so Aporia can assume the cross-account role.
In the role summary, copy the Role ARN.
Next, please provide your Aporia account manager with the Role ARN for the role you've just created.
To create a new model to be monitored in Aporia, you can call the aporia.create_model(...) API:
Each model in Aporia contains different Model Versions. When you (re)train your model, you should create a new model version in Aporia.
Each raw input, feature or prediction is mapped by default to the column of the same name in the Glue table.
By creating a feature named amount or a prediction named proba, for example, Aporia will expect a column in the table named amount or proba, respectively.
If your data format does not fit exactly, you can use Spark SQL queries to shape it in any way you want.
Next, create an instance of GlueDataSource and pass it to apr_model.connect_serving(...) or apr_model.connect_training(...):
Note that as part of the connect_serving API, you are required to specify additional 2 columns:
id_column - A unique ID to represent this prediction.
timestamp_column - A column representing when did this prediction occur.
For more information on:
Advanced feature / prediction <-> column mapping
How to integrate delayed actuals
How to integrate training / test sets
Please see the page.
