Red Hat

Watching a security fix move from pull request to closed change record without a single manual handoff is something most teams assume requires months of integration work. This demo shows it happening in one automated flow. The scenario: a CVE is flagged in a Python dependency. A Lightwell-remediated artifact drops in. Developer raises a PR, CI validates the fix, and from that moment Ansible Automation Platform runs the entire delivery pipeline. Dev deployment. Integration tests. ServiceNow work notes updating at every milestone. An approval gate that shows the approver exactly what they're signing off on, which CVE it remediates, which CI build, which change record. Load balancer drain. Production deployment. Smoke test. And if the smoke test fails, automatic rollback to the pre-deployment snapshot. The change request closes with a complete audit trail. No manual incident write-up. No chasing people across Slack and ServiceNow to close tickets. A few things worth noting: The approval prompt isn't a generic approve/reject button. It surfaces the artifact version, the CVE identifier, the CI build link, and the SNOW change number. A governed decision, not a checkbox. The AI step at the end writes the close note as contextual prose, covering the application, version, remediation status, and snapshot reference. Same pattern on the failure path, where it writes the incident note and halts instead of closing. Automation orchestrator isn't replacing your CI/CD tools or your ITSM. It's coordinating the work around them, connecting the moment CI passes to the moment production is verified and the record is closed. One remediation. One governed, automated path to production. Full demo here: https://lnkd.in/eRtERP65 Curious whether your current process closes the change record automatically or if that's still a manual step at the end. #AnsibleAutomation #RedHat #CICD #ITAutomation #DevSecOps #ServiceNow

Enterprise IT environments are more complex than ever, and skills gaps can significantly impact product quality, security, and revenue. We commissioned IDC to study the business value of Red Hat Training & Certification for large organizations using Red Hat technologies. The results show a 450% three-year ROI, breaking even in just 5 months. This structured upskilling empowers engineering teams to shift from reactive troubleshooting to proactive self-sufficiency, ultimately protecting margins. Discover how your peers are optimizing their teams in the comprehensive June 2023 IDC Business Value White Paper: https://red.ht/4wmc8O0

🚀 Ready to take your AI projects from demo to production? Discover how Retrieval-Augmented Generation (RAG) can help ground large language models in your private data with the Engineering RAG for the Enterprise e-book. Learn to design, deploy, and optimize contextually accurate generative AI applications at scale using Red Hat OpenShift AI. Get a practical framework for building robust retrieval pipelines, integrating vector databases securely, and maintaining high-throughput accuracy across hybrid cloud environments. Download your free e-book today and scale your enterprise RAG applications with production reliability: https://red.ht/3TgCTGf

What if you didn't have to overhaul your entire software stack just to patch a single vulnerability? See how Lightwell Network delivers targeted, backported fixes to help keep your production environments stable.

How do you build a responsive, reliable AI assistant for critical healthcare information? The University of North Carolina built SARHAchat on Red Hat OpenShift AI, creating an empathetic user experience with built-in safety measures. Using an open source platform gave the team greater stability and control while reducing maintenance costs and model complexity. Discover how UNC is scaling AI for future healthcare use cases: https://red.ht/4AMic63

Supporting a new AI accelerator isn’t just a hardware challenge. It’s a software validation challenge, too. In a new PyTorch blog, researchers from IBM and Red Hat share how the Torch Spyre team built on PyTorch’s CRCR framework to automate test selection, adaptation, and regression tracking across a fast-moving software stack. The goal: reduce the manual effort required to keep accelerator backends aligned with upstream PyTorch while increasing confidence that changes across the stack don’t introduce regressions. As AI infrastructure becomes increasingly heterogeneous, this kind of scalable validation will be essential. Supporting more accelerators shouldn’t mean multiplying the complexity of testing and maintenance. This work offers a practical path forward and is another example of what open collaboration across the PyTorch, IBM, and Red Hat communities can enable. Read the full blog: https://lnkd.in/eSdF5CEF

Supporting a new AI accelerator isn’t just a hardware challenge. It’s a software validation challenge, too. In a new PyTorch blog, researchers from IBM and Red Hat share how the Torch Spyre team built on PyTorch’s CRCR framework to automate test selection, adaptation, and regression tracking across a fast-moving software stack. The goal: reduce the manual effort required to keep accelerator backends aligned with upstream PyTorch while increasing confidence that changes across the stack don’t introduce regressions. As AI infrastructure becomes increasingly heterogeneous, this kind of scalable validation will be essential. Supporting more accelerators shouldn’t mean multiplying the complexity of testing and maintenance. This work offers a practical path forward and is another example of what open collaboration across the PyTorch, IBM, and Red Hat communities can enable. Read the full blog: https://lnkd.in/eSdF5CEF

Today we’re announcing OpenClaw Enterprise In collaboration with Red Hat, NVIDIA and OpenAI the OpenClaw Foundation is open sourcing a powerful enterprise control plane for persistent agents OpenClaw Enterprise is built to run on your own infrastructure and will always be free for an organization to use https://lnkd.in/gYyr8z6D

AI agents can use tools, access data, call APIs and take actions across enterprise systems. As that authority grows, security needs to extend beyond safeguards inside the model. Enterprises need a way to connect policy and risk requirements to technical controls, and to enforce those controls across the systems surrounding the agent. Red Hat is working with NVIDIA on that systems approach. The asago community is exploring how AI governance policies can translate into testable risks and technical controls. OpenShell provides mechanisms for constraining agents at runtime, while BlueField creates opportunities to place selected enforcement functions across a hardware-isolated infrastructure boundary. Together with controls across Red Hat AI, Red Hat OpenShift and Red Hat Enterprise Linux, these layers can provide independent enforcement points around what an agent is permitted to access and do. Agent security isn't just about guiding behavior. It's also about establishing and enforcing the boundaries of an agent's authority. Read the news: https://red.ht/3T3RJjl