Overview

Glue is a serverless AWS offering used for data cataloging, transformation, integration, and orchestration. The Monte Carlo integration is primarily interested in the cataloging features of the Glue Metastore. The Glue Metastore provides Monte Carlo with the databases, tables, and partition information of data stored in AWS S3 Buckets.

How does Monte Carlo use Glue?

The catalog information in Glue powers the Monte Carlo Assets view providing users with

The ability to discover and explore data assets cataloged by Glue
Table metadata type, schema, column, and partition information
Schema monitoring
Query Engines (Other data lake components rely on the Glue metastore to effectively map their queries to your data assets)

🚧
Databricks with Glue
If you are using Databricks with an external Glue catalog, please follow the Databricks documentation. Setting up Glue as a separate integration is not necessary.

What features are not enabled by a Glue integration?

Monte Carlo data lake observability is enabled by integrating the entire data lake stack. The Glue metastore is a critical piece of that stack, but does not enable the full suite of data observability features on it's own. In order to get the best coverage your data lake you will need to include a Query Engine integration to work in conjunction with Glue. Below is a list of Monte Carlo features that are not available with Glue alone.

Automated Freshness and Volume Monitoring
Field Health, Dimension Tracking and JSON Monitors
SQL, Freshness, Volume, Field Quality and Cardinality Rules
Query Logs
Importance scores
Insight Reports

How does Monte Carlo integrate with Glue?

Monte Carlo uses assumable IAM roles to reach into your Glue metastore. Using the Monte Carlo CLI, you pass the role ARN, database names, and S3 buckets you are interested in monitoring to Monte Carlo.

The easiest way to set up Glue is to use the Monte Carlo CLI exclusively (Option 1) as the CLI completely automates the infrastructure creation and policy configuration. In the case that is not possible, proceed to Option 2.

📘
Prerequisites

Permission to create IAM roles and policies in AWS

Monte Carlo CLI configured with API keys ( Please follow this guide to install and configure the CLI on your local machine)

Option 1: Use the Monte Carlo CLI [Recommended]

Generate the Glue access policy
Create an access role
Provide role information to Monte Carlo

1. Generate the Glue access policy

Run montecarlo discovery glue-policy-gen [parameters] > glue_access_policy.json with the necessary parameters. If the data account is not the same as the collector account, use --resource-aws-region and --resource-aws-profile to pass the data account profile.

$ montecarlo discovery glue-policy-gen --help
Usage: montecarlo discovery glue-policy-gen [OPTIONS]

  Generate an IAM policy for Glue. After review, output of this command can
  be redirected into `montecarlo integrations create-role` or `montecarlo
  discovery cf-role-gen` if you prefer IaC.

Options:
  --database-name TEXT         Glue/Athena database name to generate a policy
                               from. Enter '\*' to give Monte Carlo access to
                               all databases. This option can be passed
                               multiple times for more than one database.
                               [required]

  --data-bucket-name TEXT      Name of a S3 bucket storing the data for your
                               Glue/Athena tables. If this option is not
                               specified the bucket names are derived (looked
                               up) from the tables in your databases. This
                               option can be passed multiple times for more
                               than one bucket. Enter '\*' to give Monte Carlo
                               access to all buckets.

  --resource-aws-region TEXT   Override the AWS region where the resource is
                               located. Defaults to the region where the
                               collector is hosted.

  --resource-aws-profile TEXT  Override the AWS profile use by the CLI for the
                               resource. This can be helpful if the resource
                               and collector are in different accounts.

  --collector-id UUID          ID for the data collector. To disambiguate
                               accounts with multiple collectors.

  --help                       Show this message and exit.

2. Create an access role

Run montecarlo integrations create-role glue_access_policy.json. If the data account is not the same as the collector account, use --aws-profile to pass the data account profile.
The command prints a role ARN and an external id, they are used in the next section.

$ montecarlo integrations create-role --help
Usage: montecarlo integrations create-role [OPTIONS] FILE

  Create an IAM role from a policy FILE. The returned role ARN and external
  id should be used for adding lake assets.

Options:
  --aws-profile TEXT  Override the AWS profile used by the CLI, which
                      determines where the role is created. This can be
                      helpful when the account that manages the asset is not
                      the same as the collector.

  --help              Show this message and exit.

3. Provide role information to Monte Carlo

Run montecarlo integrations add-glue with the necessary parameters.

$ montecarlo integrations add-glue --help
Usage: montecarlo integrations add-glue [OPTIONS]

  Setup a Glue integration. For metadata.

Options:
  --region TEXT        Glue catalog region. If not specified the region the
                       collector is deployed in is used.
  --role TEXT          Assumable role ARN to use for accessing AWS resources.
                       [required]
  --external-id TEXT   An external id, per assumable role conditions.
  --name TEXT          Friendly name for the created warehouse. Name must be
                       unique.
  --collector-id UUID  ID for the data collector. To disambiguate accounts
                       with multiple collectors.
  --skip-validation    Skip all connection tests. This option cannot be used
                       with 'validate-only'.
  --validate-only      Run connection tests without adding. This option cannot
                       be used with 'skip-validation'.
  --auto-yes           Skip any interactive approval.
  --option-file FILE   Read configuration from FILE.
  --help               Show this message and exit.

Option 2: Use the AWS UI

1. Create an access role

Follow the steps outlined in Creating IAM Roles to create a role with the policy below, replacing the values REGION, ACCOUNT_ID, S3_ARN (of the S3 bucket(s) storing the data for your Glue/Athena tables - you can alternatively pass "*" to give access to all buckets), and DATABASE_NAME.
The role ARN and external ID should be saved to be used in the next step.

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "s3:ListBucket",
                "s3:GetBucketLocation"
            ],
            "Resource": [
                "<S3_ARN>"
            ]
        },
        {
            "Effect": "Allow",
            "Action": "glue:GetConnections",
            "Resource": [
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:catalog",
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:connection/*"
            ]
        },
        {
            "Effect": "Allow",
            "Action": "glue:GetDatabases",
            "Resource": [
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:catalog",
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:database/<DATABASE_NAME>"
            ]
        },
        {
            "Effect": "Allow",
            "Action": [
                "glue:GetTables",
                "glue:GetTable",
                "glue:GetPartitions",
                "glue:GetPartition"
            ],
            "Resource": [
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:catalog",
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:database/<DATABASE_NAME>",
                "arn:aws:glue:<REGION>:<ACCOUNT_ID>:table/<DATABASE_NAME>/*"
            ]
        }
    ]
}

2. Provide role information to Monte Carlo

Run montecarlo integrations add-glue with the necessary parameters.

$ montecarlo integrations add-glue --help
Usage: montecarlo integrations add-glue [OPTIONS]

  Setup a Glue integration. For metadata.

Options:
  --region TEXT        Glue catalog region. If not specified the region the
                       collector is deployed in is used.
  --role TEXT          Assumable role ARN to use for accessing AWS resources.
                       [required]
  --external-id TEXT   An external id, per assumable role conditions.
  --name TEXT          Friendly name for the created warehouse. Name must be
                       unique.
  --collector-id UUID  ID for the data collector. To disambiguate accounts
                       with multiple collectors.
  --skip-validation    Skip all connection tests. This option cannot be used
                       with 'validate-only'.
  --validate-only      Run connection tests without adding. This option cannot
                       be used with 'skip-validation'.
  --auto-yes           Skip any interactive approval.
  --option-file FILE   Read configuration from FILE.
  --help               Show this message and exit.

Troubleshooting

Validation error

User: arn:aws:iam::<aws_account_id>:user/<user> is not authorized to perform: <action> on resource: arn:aws:glue:<region>:<aws_account_id>:<resource>

Potential Causes

AWS policy is not attached to a Role
Policy is misconfigured. This often happens when the policy is created by hand. Try generating the the policy using Option 1 and performing a diff between your policy and the Monte Carlo CLI generated one
A resource policy (e.g. bucket policy) is denying access
Other types of policies are denying access, see AWS doc, but keep in mind that this is less likely, things like permission boundaries are not common.

Assume role error

User: arn:aws:sts::<dc_account_id>:assumed-role/<role> is not authorized to perform: sts:AssumeRole on resource: arn:aws:iam::<data_account_id>:role/<role>

Potential Causes

The role was created under the wrong AWS account. The IAM role must exist in the same account as Glue
The trust relationship account is wrong - the account in the trust relationship must be the data collector account
A trust relationship is missing (step 2 in creating IAM role docs)
The MonteCarloData tag is missing or contains a value - value must be absent (step 7 in IAM role docs)