Menka Rajput’s Post

🚨 This might change how on-call works forever. AWS just released something big on March 31, 2026. AWS DevOps Agent is now Generally Available. I have not used it yet, but what it promises is hard to ignore 👇 Imagine having an always-on operations teammate that: → Starts investigating incidents the moment an alert fires, even at 2 AM → Connects signals across your observability tools, runbooks, CI/CD pipelines, and code repositories → Does not just detect issues but suggests how to prevent them next time → Works across AWS, Azure, and on-prem environments That is what AWS DevOps Agent is aiming to be. And the early numbers are pretty impressive: 📉 MTTR reduced by up to 75% 🎯 94% root cause accuracy 🔍 80% faster investigations ⚡ 3 to 5 times faster incident resolution What stands out to me 👇 This does not feel like just another AI chatbot sitting on top of your tools. It is built to understand how your systems work, learn dependencies, and actively help with incident triage. That means less firefighting and more time building. Also, the multicloud support is a big deal. Most tools stop at AWS. This one goes beyond. I am sharing this as an update, not a recommendation yet. Really curious to see how it performs in real environments. Have you tried AWS DevOps Agent? Would love to hear your experience 👇 #AWS #DevOps #SRE #CloudEngineering #AIAgents #CloudComputing #IncidentManagement #PlatformEngineering #TechTrends #Kubernetes

AWS just made a big move for modern Ops teams: AWS DevOps Agent is now generally available. What caught my attention (from a DevOps/SRE perspective) :- 𝐅𝐚𝐬𝐭𝐞𝐫 𝐢𝐧𝐜𝐢𝐝𝐞𝐧𝐭 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐞: investigates alerts, correlates logs/metrics, summarizes “what changed” 𝐑𝐞𝐥𝐢𝐚𝐛𝐢𝐥𝐢𝐭𝐲 + 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧: suggests fixes from real operational signals 𝐖𝐨𝐫𝐤𝐬 𝐛𝐞𝐲𝐨𝐧𝐝 𝐀𝐖𝐒: handles AWS, multicloud, and on-prem 𝐄𝐱𝐭𝐞𝐧𝐬𝐢𝐛𝐥𝐞: add custom skills, build charts/reports for deeper dives My takeaway: A step toward agentic ops—but winning setups stay human-in-the-loop with least privilege + full audit trails. Ops folks: If you could deploy an ops agent today, first on incident triage, change risk review, or auto-remediation? #AWS #DevOps 

AWS is introducing a DevOps agent that can automatically revert changes when something breaks — a big shift from the usual flow where something fails, alerts fire, and engineers jump in to fix it. On paper, this sounds like a huge win: • bad deployments get rolled back instantly • broken configs are corrected before impact spreads • less manual debugging and fewer late-night fixes. But it also raises some real questions about performance and trust: • how accurately can it identify the true root cause vs just reacting to symptoms? • what happens in complex cases like partial deploys or data migrations? • can an automated rollback actually make things worse in certain scenarios? I like where this is going — but I think the real value will depend on how well it handles edge cases and whether teams can put the right guardrails around it. I’d especially be interested to hear from engineering managers — are you ready to rely on agents for decisions like this, or does it still feel too early? #DevOps #AWS #AIOps #Cloud #PlatformEngineering

🚨 AWS just dropped something big — March 31, 2026. AWS DevOps Agent is now Generally Available. I haven't used it yet — but the numbers are hard to ignore. Here's what it does 👇 Imagine an always-on operations teammate that: → Investigates incidents the moment an alert fires — yes, even 2 AM → Correlates your observability tools, runbooks, CI/CD pipelines and code repos → Doesn't just find the problem — it recommends how to prevent it next time → Works across AWS, Azure and on-prem — true multicloud support That's AWS DevOps Agent. Early customer numbers: 📉 MTTR reduced by up to 75% 🎯 Root cause accuracy at 94% 🔍 Investigation time cut by 80% ⚡ Incident resolution 3–5x faster What stands out to me: This isn't just another AI chatbot layered on top of your stack. It learns your applications, understands their relationships, and autonomously triages incidents — so your team spends less time firefighting and more time building. And the multicloud support is a big deal. Most tools stop at AWS. This one doesn't. I'm sharing this as news — I haven't personally used it yet. Have you tried AWS DevOps Agent in your team? What's your experience been so far? Drop it below 👇 I'd genuinely love to hear real feedback before exploring it further. #AWS #DevOps #AWSDevOps #CloudEngineering #SRE #AIAgents #CloudComputing #EngineeringLeadership #TechNews #Kubernetes

🚀 Big News from Amazon Web Services! On 31 March 2026, the General Availability of AWS DevOps Agent — your always-on, AI-powered DevOps teammate 🤖 In modern cloud environments, operations teams spend countless hours: 🔍 Investigating incidents 🔗 Correlating data across multiple tools 🚨 Manually triaging alerts All of this operational toil slows down innovation. 💡 AWS DevOps Agent changes the game: ✅ Proactively prevents and resolves incidents ✅ Learns your applications and their dependencies ✅ Integrates with observability tools, CI/CD pipelines, runbooks, and code repositories ✅ Works seamlessly across AWS, multi-cloud, and on-prem environments 📊 Impact (from preview users): • ⏱️ Up to 75% reduction in MTTR • ⚡ 80% faster investigations • 🎯 94% root cause accuracy • 🚀 3–5x faster incident resolution This is a massive step toward AI-driven SRE and autonomous operations. As someone working in Cloud & DevOps, this is exactly where the industry is heading — less firefighting, more innovation. 👉 Curious to see how this evolves in real-world production environments! #AWS #DevOps #SRE #CloudComputing #AIOps #Automation #Kubernetes #CloudEngineering #Innovation

AWS DevOps Agent went GA on March 31. It's not a copilot. It's an autonomous on-call engineer. When production breaks at 2 AM: → It traces the incident to the exact commit → Correlates logs, metrics, and deployment data → Posts root cause + fix to Slack in under 5 minutes → No human woke up Works across AWS, Azure, and on-prem (via MCP). Learns your team's patterns. Gets smarter over time. Deduplicates tickets. Triages severity automatically. Preview results: • 75% lower MTTR • 94% root cause accuracy • 3-5x faster incident resolution Built on Bedrock AgentCore — not a thin wrapper over an LLM. It has its own memory, policies, topology awareness, and observability stack. For lean DevOps teams in ASEAN, this is the kind of force multiplier that changes what's possible. https://lnkd.in/gtG8dnny AWS is no longer building tools. They're building teammates. #AWS #DevOps #AgenticAI #SRE #CloudOps #AIOps

🚨 AWS just dropped something that could fundamentally change how on-call works. Introducing the AWS DevOps Agent 👉 https://lnkd.in/gqw-ExVb This isn’t just another automation tool. It’s an AI-powered, always-on DevOps engineer that investigates, resolves, and even prevents incidents. (Amazon Web Services, Inc.) 💡 What makes this a big deal? 🤖 Autonomous incident response Starts investigating the moment an alert fires — no human needed to kick things off. (Amazon Web Services, Inc.) 🔍 Deep root cause analysis Correlates telemetry, code, deployments, and dependencies across your entire stack. (Amazon Web Services, Inc.) ⚡ Reduced MTTR (from hours → minutes) Early users report dramatic improvements in resolution time. (Amazon Web Services, Inc.) 🔁 Proactive prevention Learns from past incidents and recommends improvements in observability, infra, and pipelines. (Amazon Web Services, Inc.) 🔗 Seamless integrations Works with tools you already use: CloudWatch, Datadog, Splunk, GitHub, PagerDuty, and more. (Amazon Web Services, Inc.) 🧠 Why this matters (especially for SREs): We’ve spent years: Writing runbooks Tuning alerts Chasing logs across tools Doing 2 AM firefighting Now imagine: 👉 An agent that understands your entire system topology 👉 Investigates like a senior engineer 👉 Suggests (or even executes) mitigation steps This is a shift from: Reactive Ops → Autonomous Reliability Engineering ⚠️ My take: This is powerful — but also disruptive. Where does human intuition still matter? How do we validate agent decisions in high-risk systems? What does “on-call” even mean in 2–3 years? 🔥 The bigger trend: We’re moving toward agentic infrastructure — where systems don’t just alert, they act. And honestly, this might be the beginning of: 👉 “Zero-touch operations” 👉 AI-driven SLO enforcement 👉 Self-healing distributed systems Curious — would you trust an AI agent to debug your production at 2 AM? Video: https://lnkd.in/gTq6ENdJ #AWS #DevOps #SRE #AI #CloudComputing #PlatformEngineering #Observability #Innovation

💸 Companies are wasting thousands of dollars every month on AWS… without even realizing it. One of the most common hidden problems? 👉 Over-Provisioned EKS / ECS Clusters Many teams allocate: ⚙️ More CPU ⚙️ More Memory ⚙️ More Nodes …than the application actually needs. Result? 🚨 Paying for resources that are NOT being used. Example: Provisioned: 8 vCPU + 16GB RAM across 6 nodes Actual usage: 2 vCPU + 5GB RAM across 2 nodes ➡️ Nearly 70% infrastructure cost wasted 🔍 How to identify Over-Provisioning Check metrics in CloudWatch: • CPU utilization consistently below 30% • Memory utilization below 40% • Idle pods or unused services running • Fixed capacity for variable workloads If this happens → your cluster is over-provisioned. 💡 How to Optimize EKS / ECS Cost ✔ Rightsize CPU & Memory based on real usage ✔ Enable Horizontal Pod Autoscaler (HPA) ✔ Use Cluster Autoscaler for dynamic scaling ✔ Use Spot Instances (70–90% cheaper) ✔ Remove unused containers & services ✔ Monitor with CloudWatch / New Relic 📚 Skills to learn for AWS Cost Optimization Engineer role If you want to move into FinOps / Cloud Optimization, focus on: • AWS Lambda cost optimization patterns • EKS & ECS container rightsizing • RDS Aurora performance tuning • DynamoDB cost strategies • Savings Plans vs Reserved Instances • CloudWatch metrics analysis • CDK (TypeScript) infrastructure optimization • Auto Scaling strategies • Observability tools (New Relic, CloudWatch) 🎯 Real Impact Cost optimization is not just about saving money… It improves: • performance • scalability • engineering efficiency • architecture maturity Companies value engineers who can optimize both performance AND cost. 👇 Are you learning AWS or DevOps? Comment “COST” and I will share more real-world cost optimization interview questions. Let’s connect and grow together 🤝 Follow for more content on: AWS • DevOps • FinOps • Cloud Architecture • Interview Preparation #AWS #DevOps #FinOps #CloudComputing #EKS #ECS #CostOptimization #AWSCertified #CloudEngineer #TechCareers #LearningInPublic #IndiaTech #CareerGrowth 🚀

Most comparisons between AWS and Azure stop at “EC2 vs VM”… But real engineering decisions happen at the architecture level 👇 ☁️ AWS vs Azure — From a DevOps / Platform Engineering Perspective 🔹 1. Identity & Access (Critical for Security) AWS → IAM (policy-based, fine-grained control) Azure → Azure AD + RBAC (directory-first approach) 👉 AWS = resource-centric permissions 👉 Azure = identity-centric permissions 💡 In large enterprises, Azure AD integration is a game changer. 🔹 2. Networking Philosophy AWS → VPC is isolated by default (you design everything) Azure → VNet integrates more natively with services 👉 AWS gives more control 👉 Azure gives more abstraction 🔹 3. Kubernetes (EKS vs AKS) EKS → More control, but more setup overhead AKS → More managed, faster to get started 💡 In production: EKS = flexibility AKS = speed 🔹 4. DevOps Ecosystem AWS → Modular (CodePipeline, CodeBuild, etc.) Azure → Integrated (Azure DevOps end-to-end) 👉 AWS = “build your stack” 👉 Azure = “use the platform” 🔹 5. Multi-Account vs Subscription Model AWS → Multi-account strategy (best practice for isolation) Azure → Subscription + Management Groups 💡 AWS handles scale with account boundaries 💡 Azure handles scale with hierarchy 🔹 6. Infrastructure as Code Reality Both support Terraform, BUT: 👉 AWS → More community modules, faster updates 👉 Azure → Better native ARM/Bicep integration 🔹 7. Observability & Monitoring AWS → CloudWatch (powerful but fragmented) Azure → Azure Monitor (more unified experience) 🚀 What Senior Engineers Focus On It’s NOT: ❌ “AWS vs Azure” It’s: ✔️ Designing fault-tolerant systems ✔️ Managing blast radius (multi-account / subscriptions) ✔️ Securing identities, not just resources ✔️ Automating everything (IaC + CI/CD) ✔️ Observability & incident response 💡 Final Thought Cloud is just a tool. The real skill is: 👉 Understanding distributed systems 👉 Designing for failure 👉 Automating at scale If you’re learning DevOps today, focus on concepts over cloud vendors — that’s what makes you truly cloud-agnostic. #AWS #Azure #DevOps #PlatformEngineering #Kubernetes #CloudArchitecture #SRE #TechIns

Stop just learning AWS services. Start building real DevOps workflows. 🚀 There’s a massive gap between understanding a service and actually running it in a live production environment. Learning EC2, S3, or Lambda is great. But connecting compute, networking, security, and automation into one resilient system? That’s where the real engineering happens. I’ve structured an AWS DevOps Digital Guide that bridges this exact gap—taking you from foundational concepts to real, CLI-driven production scenarios. Here’s a snapshot of the workflows we cover: 🔒 Security & Networking: IAM least privilege, VPC design, and private subnets. ⚙️ Automation & CI/CD: End-to-end pipelines using CodePipeline, CodeBuild, and CodeDeploy. 🏗️ Infrastructure as Code: Making setups repeatable with CloudFormation and CDK. 📦 Modern Compute: Containerizing with ECS/EKS and building serverless architectures. 📊 Day-2 Ops: CloudWatch monitoring, cost optimization, and Auto Scaling under load. Why does this matter? Because modern cloud engineering isn't a vocabulary test of AWS tools. It’s about building secure, automated, and cost-effective systems that solve real business problems. Whether you're prepping for your next big interview or building out enterprise infrastructure, you'll find value here. #AWS #DevOps #CloudComputing #CloudEngineering #CICD #AWSDevOps #DevOpsEngineer 👇 Question for the network: What’s the toughest AWS service you’ve had to implement in a real-world scenario? Let me know in the comments!