PJ Kirner

The guide demonstrates how LoxiLB can work with multiple CNIs in a Multus environment, offering load-balancing capabilities for secondary network interfaces alongside the primary Kubernetes network. More: https://lnkd.in/gpB-kqNB

87

If you're running Kubernetes clusters, the Metrics Server is a critical tool for tracking CPU and memory usage. Unlike Prometheus, which is used for long-term monitoring, Metrics Server focuses on real-time data that can help with auto-scaling and improving performance. In our latest guide, we dive into the Kubernetes Metrics Server—what it is, why it’s essential, and how to set it up. Check it out to enhance your Kubernetes monitoring strategy! 🔗 https://lnkd.in/gw9-cYW6 #Kubernetes #DevOps #Kubecost

9

Bringing Together Splunk Observability & AppDynamics for Unified Visibility Thursday, September 26, 2024  |  12PM SGT / 2PM AEST End-to-end, unified visibility is key to quickly troubleshooting and ensuring that, when issues do occur, your business and end user’s experience are affected as little as possible. But how do you truly achieve end-to-end visibility? By embracing full-stack observability and the power of the integrated Splunk Platform, Splunk Observability Cloud, AppDynamics, and Splunk IT Service Intelligence (ITSI). Join this session to explore the ins and outs of this integration, how it can help you and your team gain end-to-end visibility and accelerate troubleshooting, and how other customers have achieved real-life success. In this webinar you will learn: - How Splunk Platform, Observability Cloud, ITSI, and AppDynamics are currently being integrated - Why end-to-end visibility is critical and how it helps accelerate troubleshooting - How to implement this integration and what software components are involved - Real-life success stories showcasing the business outcomes and ROI achieved by other customers Use the link below to register: https://lnkd.in/gAZeiHhh

10

Transform your IAC validation process with VR's OPA Policy Generator. This intuitive tool simplifies the creation of secure, compliant policies for Terraform AWS-based infrastructure, cutting through the complexity of OPA's syntax and making it accessible for all developers and security experts. 🔒✨ #VRModuleSpotlight #OPA #PolicyGenerator #IAC #AWS #AWSPartner #APNProud Amazon Web Services (AWS) AWS Partners

6

With the rise of quantum computer adoption comes the potential risks and impact on cybersecurity. It is imperative for enterprises to take a proactive approach to quantum-resilient technologies. Learn how organizations can leverage Quintessence Labs' cybersecurity options at our #datacenters to strengthen defences against cryptographic threats posed by quantum computing.  https://eqix.it/4el2odZ 

1

National Institute of Standards and Technology (NIST) released 3 new PQC standards today, representing a major milestone in keeping information on the Internet secure and confidential. See how Google is using PQC, and how orgs can adopt these new standards - https://lnkd.in/eSR_rqAr

102

1 Comment

🔍 Critical ML Vulnerabilities Uncovered by JFrog! JFrog has identified 22 vulnerabilities in open-source ML projects, exposing risks like remote model hijacking and unauthorized privilege escalation. As AI adoption rises, the need for stronger security has never been clearer. 🔗Learn more: https://lnkd.in/ghvUxY9Q #AI #Cybersecurity #ISBNews

16

Nested virtualization or checkpoint restore? -Which is better for running your Spot EC2 workloads? -Many customers I’ve met with don’t fully understand the differences and trade-offs. Let’s start with definitions: Checkpoint/Restore Checkpoint/restore saves an application’s state (memory, disk, and process info) at specific intervals, allowing it to restart from the last saved state after a Spot instance interruption. Nested Virtualization Nested virtualization runs a virtual machine (VM) inside an EC2 Spot instance (another VM), decoupling workloads from the host and enabling full VM migration during interruptions. Here are the tradeoffs: -Both work for stateful applications, the hardest workloads to reliably run on Spot. -Nested virtualization has a higher resource overhead, so your total realizable benefit will be more limited. -Some of the additional overhead of nested virtualization can be reduced if you modify the Linux kernel, but do you really wanna go there? -On the other hand, apps such as databases with lots of external dependencies can be trickier with checkpoint restore. HPC apps tend to perform really well with a C+R approach. -Checkpoint restore has less dependencies on the cloud platform provider itself, so your Spot workloads are more portable compared to the specific ways nested virtualization needs to interact with the hypervisor layer unique to each cloud provider. -Integration with complex tech stacks is also easier with a checkpoint restore software library approach. I’ve heard from several customers that services like AWS Batch and 3rd party + custom job schedulers can be hard to integrate with nested virtualization approaches. Hope this was helpful. Would love to hear the experiences both good and bad from others who have tried these approaches on clouds like AWS, GCP, and others. #AWS #HPC #Genomics #EDA #FinancialServices #FinOps

7

Check out this video where I explain where DagKnows fits in a typical incident response and remediation workflow. If you see a need for a tool to perform initial diagnostic steps and enrich your alerts or even remediate problems automatically then this tool is for you! In the next part, we go over the actual demo. Stay tuned.

19

JFrog and GitHub have strengthened their alliance with a unified dashboard, streamlining vulnerability tracking from source code to deployed binaries. Now, DevOps teams can seamlessly navigate between GitHub Actions Workflows and JFrog Artifactory repositories, with visibility into security scans, permissions, and identities. This bi-directional integration makes it easier to prioritize vulnerabilities and ensure software is secure at every step. https://bit.ly/3TnvjGo #swampUP2024 #DevSecOps

2

Last call for tomorrow's Hands-on Labs! Guided by a live instructor, you'll get access to a pre-configured Kubernetes cluster and explore key features such as effortless volume provisioning, overcoming CSI limitations, mitigating “noisy neighbor” issues, and automated capacity management. Spots are filling up fast, so register now to experience Portworx in action! https://lnkd.in/g93jbu5i #kubernetes #portworx #purestorage #csi #containers #cloudnative

8

🕵️ Unlock the secrets of troubleshooting containerized apps 👀 Attend our k8s application failure workshop! In this hands-on workshop, we’ll dive into real-world scenarios and learn essential techniques to: ✅ Debug unexplained OOMKills ✅ Resolve mismatched Kubernetes configurations ✅ Overcome connection throttling issues ✅ Unblock stubborn threads Don’t miss this chance to elevate your troubleshooting game! Sign up now: https://ln-s.io/4eyBGPb cc: Prequel.dev #DevOps #Kubernetes #k8s #Observability #SRE #CloudInfrastructure

15

1 Comment

Exciting developments in the world of Large Language Models (LLMs)! A new open-source framework called RouteLLM is making waves in cost-effective LLM deployment. As someone deeply involved in AI and machine learning, I'm thrilled to share my thoughts on this groundbreaking research. RouteLLM, developed by researchers including Isaac Ong and Amjad Almahairi, addresses a critical challenge in LLM deployment: balancing performance and cost. As we know, more capable models often come with higher price tags, creating a dilemma for real-world applications. RouteLLM offers a clever solution by intelligently routing queries between stronger, more expensive models and weaker, cheaper ones. What sets RouteLLM apart is its novel approach using preference data for training routers. By leveraging comparisons of response quality between different models, the framework learns to make informed routing decisions. The researchers have explored various router architectures, including similarity-weighted ranking, matrix factorization, and both BERT and causal LLM classifiers. The results are impressive. In benchmarks like MT Bench, MMLU, and GSM8K, RouteLLM achieves significant cost reductions (up to 85% on MT Bench!) while maintaining 95% of GPT-4's performance. This could be a game-changer for organizations looking to deploy LLMs at scale without breaking the bank. I'm particularly intrigued by the data augmentation techniques explored in this research. The team demonstrated that even a small amount of task-specific data can dramatically improve routing performance. This scalable approach could pave the way for more efficient and adaptable LLM deployments across various domains. Another standout feature is the framework's ability to generalize to new model pairs without retraining. This flexibility is crucial in the rapidly evolving landscape of LLMs, where new models are constantly being released. As an open-source initiative, RouteLLM provides a valuable resource for the AI community. The researchers have released their code, datasets, and pre-trained routers, enabling others to build upon their work and potentially accelerate progress in this area. Looking ahead, I see immense potential for RouteLLM in practical applications. From chatbots and virtual assistants to content generation and data analysis tools, this framework could help organizations optimize their LLM usage and deliver high-quality AI-powered services more cost-effectively. However, it's important to note that while RouteLLM shows promising results, there's still room for improvement. As the field evolves, we may see even more sophisticated routing techniques emerge, perhaps incorporating multi-model routing or dynamic adaptation based on real-time performance metrics. https://lnkd.in/dtmUjRfq

6

Want to control the Kubernetes environments cost? Great blog by Chip Zoller where he explains how organizations can use two best-in-class tools, Kubecost for cost visibility and monitoring, and Kyverno for policy-as-code, to realize some massive savings. https://hubs.ly/Q02Hlntc0 #kubernetes #devops

11

The modern IR guide looks more and more like this: Step 1: Find out where MFA is not enabled Step 2: Find out why logs were only partially enabled Step 3: Find out if an existing leak on BreachForums is present Step 4: Find out if which user clicked on a fake Google Ad that make them install an infostealer Step 5: Fill in your 8-K Form

6