When using AWS IoT Events, you now have the option to define actions to invoke AWS Lambda functions, publish messages to the Amazon Simple Queue Service (SQS) queue or an Amazon Kinesis Data Firehose delivery stream, and republish messages to IoT Events. Previously, you could only define actions to publish messages to SNS and MQTT. These expanded actions make it easier to build monitoring applications that help you quickly understand the state of your devices by providing more options to process messages created by IoT Events.

You can now specify additional worker types when you use AWS Glue development endpoints. An AWS Glue development endpoint is a serverless Apache Spark environment that you can use to develop, debug, and test your AWS Glue ETL scripts in an interactive manner.

AWS Glue has updated its Apache Spark infrastructure to support Apache Spark 2.4.3 (in addition to Apache Spark 2.2.1 ) for ETL jobs, enabling you to take advantage of stability fixes and new features available in this version of Apache Spark.  

Amazon S3 announced support for percentiles on Amazon CloudWatch Metrics. This feature allows customers to visualize and alarm on p90, p95, p99, p99.9 or any other percentile (including p100) of an S3 request metric. This provides customers with more granularity about their request patterns on S3 and helps them observe and diagnose anomalies in request patterns on S3. 

Amazon GuardDuty is now available in the AWS Asia Pacific (Hong Kong) region. You can now continuously monitor and detect security threats in the Asia Pacific (Hong Kong) region to help protect your AWS accounts and workloads.

AWS Secrets Manager now supports VPC endpoint policies, making it easier for you to restrict egress of secrets from your Amazon VPC. When you create a VPC endpoint for Secrets Manager, you can attach an endpoint policy to define the Secrets Manager actions that can be performed, the secrets these actions can be performed on, the IAM users or roles that can perform these actions, and the accounts that can be accessed via the VPC endpoint. 

The Temporary Queue Client for Amazon Simple Queue Service (SQS) is now available. The client supports common messaging patterns such as request-response, and helps you save development time and deployment costs when creating application-managed temporary queues.

Starting today, you can launch Amazon EC2 Spot Instances running Red Hat Enterprise Linux (RHEL) on base Red Hat Enterprise Linux Images (AMIs). Previously, only customers with existing Red Hat Enterprise Linux Premium subscriptions could launch Amazon EC2 Spot Instances running RHEL (i.e. the bring your own license model). Spot Instances can now be launched through RHEL’s basic subscription model and are included in the hourly Spot Instance price.

Starting today, you can access custom-generated Amazon EC2 resource optimization recommendations in AWS Cost Explorer. These recommendations identify idle and underutilized instances across your accounts and regions. To generate these recommendations, AWS analyzes your historical EC2 resource usage, your Amazon CloudWatch metrics, and your existing reservation footprint to identify opportunities for cost savings (e.g., by terminating idle instances or downsizing active instances to lower-cost options). For example, if your m5.2xlarge has a maximum utilization of 20% over the last 14 days, AWS may recommended downsizing that instance to m5.xlarge or m5.large and show you how much you can save based on your usage and your applicable m5 family reservations.

AWS Backup now provides customers a more seamless way to manage their backups, by automatically copying tags from their resources to their backups. For customers using tags to manage their AWS resources, AWS Backup will enable them to more effectively search for source resources or conduct billing for their backups. See AWS Tagging Strategies for tagging best practices.