- Newest
- Most votes
- Most comments
Yes, this is expected behavior from the AWS Glue managed runtime and no customer action is required on your part.
AWS Glue version 5.0 uses AWS SDK for Java version 2.29.52 as part of its managed runtime environment. However, the AWS SDK for Java 1.x references you're observing in CloudWatch Logs and CloudTrail are from AWS Glue's internal service components, not from your job code or dependencies.
Specifically, the CloudTrail entries showing "aws-internal/3 aws-sdk-java/1.x" for CloudWatch Logs operations (CreateLogGroup/CreateLogStream) under the AWSGlueServiceRole are part of AWS Glue's internal logging infrastructure. These are service-managed operations that occur during the GlueBootstrapping phase to set up logging for your job execution.
Since your job script uses boto3 (the Python SDK) and you haven't added any custom JARs or dependencies that would introduce AWS SDK for Java 1.x, the SDK 1.x usage is entirely within AWS Glue's managed service layer. AWS is responsible for upgrading these internal components as part of the service's evolution.
The AWS Health notification you received is likely part of AWS's broader communication about SDK 1.x deprecation, but in this case, the remediation responsibility lies with the AWS Glue service team, not with you as a customer. Your Glue jobs running on version 5.0 are already using the supported runtime components for customer-facing functionality.
Sources
Migrating AWS Glue for Spark jobs to AWS Glue version 5.0 - AWS Glue
AWS Glue versions - AWS Glue
Relevant content
- asked 5 months ago
- AWS OFFICIALUpdated 9 months ago
