Custom Model Logging Debugging

• membership_id (required): The Clean Rooms membership ID
• region (required): The AWS region
• trained_model_arn (optional): ARN of the specific trained model or inference job

You MUST ask for all required parameters upfront.

Steps

1. Validate AWS Credentials and Region

• aws sts get-caller-identity
• Inform the user about the AWS account and region being used

2. Check Trained Model or Inference Job Status

Determine the resource type by inspecting the ARN: if it contains trained-model-inference-job, use the inference job call; otherwise use the trained model call. If ResourceNotFoundException, try the other.

• aws cleanroomsml get-trained-model --membership-identifier ${membership_id} --trained-model-arn ${trained_model_arn} --region ${region}
• aws cleanroomsml get-trained-model-inference-job --membership-identifier ${membership_id} --trained-model-inference-job-arn ${trained_model_arn} --region ${region}

If no ARN provided, list recent resources:

• aws cleanroomsml list-trained-models --membership-identifier ${membership_id} --region ${region}
• aws cleanroomsml list-trained-model-inference-jobs --membership-identifier ${membership_id} --region ${region}

If multiple returned, present the list and ask the user to confirm which to investigate.

Extract: logsStatus, logsStatusDetails, configuredModelAlgorithmAssociationArn, job status.

If configuredModelAlgorithmAssociationArn is not in the response, use: aws cleanroomsml list-configured-model-algorithm-associations --membership-identifier ${membership_id} --region ${region}. If multiple associations are returned, present the list and ask the user to confirm which one is relevant to the resource under investigation.

3. Check Configured Model Algorithm Association Privacy Configuration

• aws cleanroomsml get-configured-model-algorithm-association --membership-identifier ${membership_id} --configured-model-algorithm-association-arn ${configured_model_algorithm_association_arn} --region ${region}
• Check privacyConfiguration.policies for:
  • trainedModels.containerLogs with allowedAccountIds (for training)
  • trainedModelInferenceJobs.containerLogs with allowedAccountIds (for inference)
• You MUST verify the customer’s account ID is included in allowedAccountIds
• If containerLogs is empty/missing, flag this as a likely root cause — but you MUST continue through all remaining steps before generating the diagnosis, as multiple issues may exist simultaneously
• Explain that logging is configured in CreateConfiguredModelAlgorithmAssociation, NOT CreateTrainedModel

4. Check ML Configuration

• aws cleanroomsml get-ml-configuration --membership-identifier ${membership_id} --region ${region}
• Extract defaultOutputLocation.roleArn — this role publishes logs
• If no ML Configuration exists (ResourceNotFoundException), flag this as a root cause — the user must create one via PutMLConfiguration. Skip Step 5 (role permissions cannot be checked without a role ARN) and continue to Step 6, as multiple issues may exist simultaneously.

5. Check ML Configuration Role CloudWatch Permissions

• aws iam get-role --role-name ${role_name}
• aws iam list-role-policies --role-name ${role_name}
• aws iam list-attached-role-policies --role-name ${role_name}
• For each inline policy: aws iam get-role-policy --role-name ${role_name} --policy-name ${policy_name}
• For each attached managed policy: aws iam get-policy --policy-arn ${policy_arn} then aws iam get-policy-version --policy-arn ${policy_arn} --version-id ${version_id}
• Required permissions: logs:CreateLogGroup, logs:CreateLogStream, logs:PutLogEvents on arn:aws:logs:*:*:log-group:/aws/cleanroomsml/*
• Also check cloudwatch:PutMetricData (requires "Resource": "*") for training metrics
• Trust policy must allow cleanrooms-ml.amazonaws.com

6. Check CloudWatch Log Groups

• aws logs describe-log-groups --log-group-name-prefix /aws/cleanroomsml/TrainedModels --region ${region}
• aws logs describe-log-groups --log-group-name-prefix /aws/cleanroomsml/TrainedModelInferenceJobs --region ${region}
• If groups exist, check streams using the discovered log group name:
  • aws logs describe-log-streams --log-group-name ${log_group_name} --order-by LastEventTime --descending --max-items 5 --region ${region}
• If groups don’t exist, this may indicate missing logs:CreateLogGroup permission. Cross-reference with logsStatus and privacy config.

7. Log Publishing Status Interpretation

Only if logsStatus is PUBLISH_FAILED:

• Each member sees their own logsStatus based on their account’s log publishing
• Do NOT suggest checking other accounts
• Accounts without an ML Configuration role show “Failed to publish logs as no ML Config role is set”
• Direct customer to check their ML Configuration role permissions (Step 5)

8. Generate Diagnosis

Identify root cause (typically: missing privacy config, missing CloudWatch permissions, missing ML Configuration, or missing trust policy). Provide exact fix with CLI commands. Reference ML roles docs (opens in a new tab) and LogsConfigurationPolicy API (opens in a new tab). Note that a new job must be run after fixing — existing failed jobs won’t retroactively publish logs.

Expected output format: