Issue #52: July 11-July 17
SUMMARY: New AWS Cloud WAN service. Microsoft Sentinel Threat Monitoring for SAP highly available. Amazon Redshift Serverless new capabilities.OReilly launches Cloud Labs.
Now build, manage, and visualize global networks with AWS Cloud WAN
The new service will let users build, manage, and monitor a unified global network that connects on-premises branch offices, data centers, and VPCs across the AWS global network.
Google Cloud also introduces its Arm-based VM offering
Became the latest player to join the ARM bandwagon with its Tau T2A AMs to include Arm-based Altra CPUs.
Amazon Redshift Serverless with new capabilities, now generally available
Executes queries by automatically provisioning and scaling capacity by workload demand; also lets you create multiple serverless endpoints per AWS account and region.
Making Microsoft Sentinel Threat Monitoring for SAP highly available with AKS
To achieve better SLA or manageability than a single Docker instance, deploy Microsoft Sentinel Threat Monitoring for SAP agent into an Azure Kubernetes Service cluster. See how to do it step-by-step
Microsoft previewing Migration tools for Azure Monitor Agent
Users must migrate from Log Analytics agent (MMA or OMS) to Azure Monitor Agent before August 2024; agent migration tools will ease this by automating the migration of agents along with built-in policies in cases where no no additional solutions or services are involved.
O’Reilly announces launch of Cloud Labs to meet high demand for cloud skills
Will provide users with temporary access to a live cloud account with expert-guided instruction through its learning platform. Currently supports Azure, but AWS and Google Cloud to follow soon.
AWS announces Log Anomaly Detection and Recommendations for Amazon DevOps Guru
Will find anomalies throughout relevant logs within an app, and provide targeted recommendations to resolve issues inside the DevOps Guru dashboard. Here’s how to use it.
Increased limits for Amazon SageMaker Automatic Model Tuning
Now run up to 250 more training jobs as part of a single tuning job; explore more hyperparameter combinations, optimize the tradeoff between wall-clock time, predictive performance, and overall cost as a result.