Tobias Schmidt’s Post

There are now tools for dissecting AWS session tokens!

We are excited to announce the release of our latest blog post exploring enhanced observability for AWS Trainium and AWS Inferentia through a new integration with Datadog. Co-authored by industry leaders from Datadog, this informative piece delves into how the integration leverages AWS Neuron to provide comprehensive monitoring of resource utilization, model execution performance, and real-time infrastructure health. As machine learning workloads scale, the need for effective observability becomes paramount. The integration allows users to easily track key performance metrics and gain insights that facilitate efficient training, optimize resource usage, and enhance overall system performance. To learn more about how this integration can empower your machine learning operations and drive high-performance outcomes, read the full article here: https://ift.tt/MAmhSW9.

As we head into Q3 and Q4, and a rush of features being designed, implemented and rolled out, remembering these fundamental learnings discussed in Mike, Alec and Becky's talk in 2023 re:Invent is crucial. Avoid modal behaviors or test them enough, have bounded queues, have mechanisms to reduce blast radius, proper error classification for signal vs noise, reduce retries or use SDKs with jitter and put backpressure to reduce impact on lower layers. Good reminders to relearn, and apply for better Customer Experience, Resilience, MTTR and avoiding outages altogether for the whole ecosystem! Also, my 9s are better than yours, perhaps doesn't help since we live in an ecosystem of microservices/dependencies, where availability and MTTR matters as a whole! #aws #resilience #cloudarchitecture https://lnkd.in/gFuQ7mY8

At AWS re:Invent today, we unveiled a significant update to our AWS integration! We set out to make vast improvements for a more secure, efficient, and user-friendly experience. These include: ✅ Secure data transferring over AWS's internal network ✅ Enhanced filtering for #CloudWatch metrics ✅ Efficient resource data collection with AWS Config Learn more about our robust #AWS integration experience: https://lnkd.in/gWrGP9GM #AWSreInvent #Lambda #DevOps #observability

Thank you to everyone who joined our AWS Compute Performance Engineering session! 🚀 Here are the top 3 insights we discussed: - #Benchmarking and #continuous monitoring are crucial for #optimization: These practices help ensure your applications run efficiently and meet performance requirements. - AWS offers great services like #Graviton, #Karpenter, and #Spot Instances: These tools help you scale and rightsize your infrastructure, providing flexibility and cost savings. - #Cost and #Performance are key for cloud #efficiency: Optimizing both ensures you get the most out of your AWS environment, especially when using containers for modern applications. Let's keep pushing the boundaries of what's possible with #AWS! #CloudComputing #PerformanceEngineering #Containers #DevOps #CloudOptimization #TechInnovation #AWSGraviton #EfficientCompute Watch the full session here:

Let's do AWS today Who can explain this?

🚀Auto Scaling is now available for AWS Glue interactive sessions Managing resources is now a thing of the past. With this new feature, AWS Glue interactive sessions automatically scale up or down workers based on actual usage, allowing you to focus more on building and less on infrastructure. Read more here: https://lnkd.in/gq2h9GGp Alona Nadler , Kinshuk Pahare , Zachary Mitchell , Saroj Yadav , Mohit Saxena , Nitin Bahadur , William Vambenepe , Keerthi Chadalavada , Xiaoxi Liu , Christopher Kha , Noritaka Sekiyama , Santosh Chandrachood

Quickly gain context when troubleshooting AWS Lambda performance issues and understand key health metrics with our useful cheatsheet: https://lnkd.in/ehfPiGpK

Great content for any AWS enthusiast looking to dive deeper into cloud infrastructure and services.

AWS Notes Consider a REPOST If It Is Useful