Manoj K.’s Post

Hot take: DevOps in 2026 is barely recognizable from what it was 3 years ago. 🔥 We used to argue about CI/CD pipelines and Dockerfiles. Now we're talking about self-healing infrastructure, AI agents writing Terraform, and pipelines that fix themselves before you even get the alert. A few things that are genuinely reshaping the space right now: → AI is inside the pipeline — not just assisting devs, but making release decisions, detecting anomalies, and rolling back deployments autonomously → Platform Engineering is eating DevOps — Internal Developer Platforms (IDPs) are becoming the default. Your team shouldn't be rebuilding the same CI scaffold from scratch every project → FinOps is now a DevOps concern — cloud bills don't lie. Cost guardrails are being baked directly into pipelines → GitOps is maturing fast — 64% adoption last year, and teams using it are reporting significantly better reliability and rollback speed → DevSecOps by default, not by afterthought — security is shifting from "we'll fix it in prod" to being enforced at the pipeline level with AI-audited checks The "move fast and break things" era is officially over. 2026 is about moving fast AND keeping things standing. 🏗️ What trend are you most focused on right now? Drop it in the comments 👇 #DevOps #PlatformEngineering #CloudNative #DevSecOps #Terraform #GitOps #AIOps

DevOps is the engine, but AIOps is the autopilot. Scaling production manually is no longer a sustainable strategy. Here is the breakdown of how traditional DevOps is evolving into AI-driven engineering: 1. CI/CD vs. Intelligent Pipelines - DevOps: Standardized GitHub Actions & Jenkins flows for delivery. - AIOps: Self-optimizing deployments that learn from past build failures. 2. Monitoring vs. AI Observability - DevOps: Setting manual thresholds in Prometheus & Grafana. - AIOps: Predictive anomaly detection using ML models to spot issues before they happen. 3. Manual Triage vs. Root Cause Analysis (RCA) - DevOps: SREs digging through logs during a production incident. - AIOps: AI agents identifying the exact code commit or config change causing the lag. 4. Cloud Ops vs. FinOps Automation - DevOps: Using Terraform for static infrastructure and resource allocation. - AIOps: Real-time cost optimization and dynamic scaling based on LLM-driven traffic patterns. DevOps builds the rails; AIOps drives the train at scale. #DevOps #AIOps #CloudComputing #MLOps #AWS #Linux #Docker #Kubernetes #Terraform #Git #Automation #SRE # 👍✌

For anyone exploring what “agentic SRE” actually looks like in practice, this is worth a look. Jason Mimick from Amazon Web Services (AWS) just shared a toolkit that brings together the building blocks of autonomous operations and putting it into practice and delivering something you can actually run and experiment with: repo: https://lnkd.in/gikpedDy What I like about this is it’s not theoretical. It reflects the shift we’re seeing across the industry: • moving from dashboards to investigation workflows • from reactive troubleshooting to AI-assisted root cause • from isolated signals to system-wide context This is the same direction AWS is pushing with AWS DevOps Agent and agentic operations, where systems can detect, analyze, and recommend actions across complex environments. If you’re working in: SRE / DevOps Platform engineering Dynatrace Observability or building with Bedrock / agent frameworks I’d recommend taking a look and seeing how it fits into your workflows. Curious who’s already testing things like this in real environments vs just exploring. #aws #devops #sre #observability #ai #platformengineering #dynatrace

I was good with Kubernetes. Now I’m super efficient and insanely productive with Kubernetes. What changed? Not another certification. Not another tool. Just one thing: configuring a Kubernetes MCP Server with LLM integration. To test this, I even tried running a pod with an intentional error - just to see how it behaves. Instead of manually debugging: I simply asked what’s wrong The AI identified the failing pod Explained why it failed (not just logs, but actual root cause) Even pointed out architecture mismatch before the container could start That’s when it really clicked. Instead of digging through YAMLs, events, and kubectl commands, I now: Ask natural language questions to understand cluster state Debug issues faster with contextual insights Generate manifests, fixes, and optimizations instantly Automate repetitive ops workflows without scripting everything from scratch It feels like moving from CLI-driven operations → intent-driven operations. This isn’t about replacing DevOps engineers. It’s about amplifying what we already know. The real shift is: From “How do I debug this?” To “Tell me what’s wrong and fix it.” If you’re already comfortable with Kubernetes, this is your next leap. Not just working harder… but working smarter. Curious to see where this goes next — especially with AI agents becoming part of the DevOps stack. #Kubernetes #DevOps #PlatformEngineering #AI #MCP #CloudNative #Productivity #Automation

New blog post from Alex Moore on Lets Do DevOps: Agentic AI Email Bot on AWS Bedrock AgentCore: Building the New article: Thread, Sending the Reply, and Operating in Production (2/2) Focuses on what happens after the email hits the worker. Rebuild the full thread, run the agent with full context, convert markdown into email safe HTML, and send replies without leaking to external recipients. Also covers attachments, error handling, and reducing Docker build time from about 21 minutes to 4. And the code is now open sourced!! Link in the article. Link to article in comments.

𝗜 𝗷𝘂𝘀𝘁 𝘀𝗵𝗶𝗽𝗽𝗲𝗱 𝘁𝗵𝗲 𝗗𝗲𝘃𝗢𝗽𝘀 𝘀𝗲𝘁𝘂𝗽 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗷𝗼𝗯𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗮𝗴𝗲𝗻𝘁 𝗯𝘂𝗶𝗹𝘁 𝗼𝗻 𝘁𝗵𝗲 Cloudflare 𝗘𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺. For a while, I was deploying manually. Every change I made = manual wrangler deploy command. That's Slow, Error-prone and doesn't scale. However, I set up a proper CI/CD pipeline that: -> Automated tests on every push -> Staging environment for previewing changes first -> Production only updates if all checks pass -> GitHub Actions handles everything Why does this matter? Because I'm not just building a project, I'm building a system that scales. When jobresearchagent grows to 10,000 users, I can't be manually deploying at 2 AM, hoping nothing breaks. This DevOps setup is one of those foundational pieces. Code quality gates before production. Automated monitoring for errors. Everything is logged, auditable, and predictable. Next: RAG system for the application. Building the retrieval layer so the AI agent has access to real, up-to-date information. If you're building AI products, Large applications, or scaling applications, this pipeline matters more than you think. The Video shows the full breakdown of this Implementation. Check out jobresearchagent.com to learn more #DevOps #AI #Engineering #Cloudflare #BuildingInPublic #jobresearchagent

Over the past few years working with Kubernetes in production, one thing has remained constant: Operating it is still harder than it should be. Even for experienced engineers, debugging an issue often means: jumping between kubectl, logs, and events opening Grafana for metrics checking alerts in another tool trying to correlate everything manually It’s powerful — but not intuitive. That’s the problem I wanted to solve. 🚀 Introducing Klariq — an AI-native control plane for Kubernetes The idea is simple: Instead of navigating multiple tools and commands, you can just ask: 💬 “Why is my service failing?” Klariq: analyzes pods, logs, and events correlates metrics and alerts identifies the root cause suggests a fix (and can even generate the YAML) All in one place. 💡 What makes Klariq different Natural language interface for cluster operations Real-time context from Kubernetes and Prometheus Local-first design — runs with your own infrastructure and LLM (no data leaves your environment) Built for engineers — not just dashboards, but actionable insights ⚙️ What I’ve been using it for Debugging incidents without switching between tools Generating Kubernetes manifests faster Understanding cluster behavior more clearly Exploring cost and security insights in one place 🧠 Why I built this Kubernetes isn’t getting simpler — but our interfaces can. We’ve moved from: SSH → kubectl kubectl → dashboards The next step is: 👉 from dashboards → intelligent, conversational control planes Klariq is my take on that. It’s still early, but already saving me a significant amount of time in day-to-day operations. 👉 https://klariq.app 👉 https://klariq.app/demo I’d really appreciate your feedback — especially from anyone working with Kubernetes in production. #kubernetes #devops #platformengineering #cloud #ai #opensource

Kubernetes didn’t really change. The workloads did. Core infra, DevOps, AI/MLOps. Kubernetes has been quietly keeping up while the job around it got more complex. If I had to pick a stack I actually trust heading into 2026, this would be it: Foundation • Docker container runtime • Helm packaging and releases • Terraform infrastructure as code DevOps • Argo CD GitOps that works at scale • Prometheus still the backbone for metrics • KEDA event driven scaling done right Security • Vault secrets without guesswork • OPA policy you can reason about • Cilium networking and security together AI / MLOps • Kubeflow ML pipelines on Kubernetes • KServe model serving that fits the ecosystem • Ollama practical LLM inference Operations • Crossplane platform engineering without glue code • Inference Gateway routing LLM traffic sanely No hype. No “ultimate stack.” Just tools that hold up when things get messy. GitHub links in the first comment. #Kubernetes #DevOps #MLOps #AI #CloudNative #PlatformEngineering

🚨 Kubernetes is overrated… Yes, I said it. People think it’s just about containers. But that mindset is already outdated. 👉 Kubernetes was never the product. 👉 It’s the platform to build everything on top of. Kubernetes was never the end goal. It’s the 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 👇 And the real game starts here 👇 ⚙️ Kubernetes + Docker → Container Orchestration 📦 Kubernetes + Helm → Package Management 🏗️ Kubernetes + Terraform → Infrastructure Provisioning 🔁 Kubernetes + ArgoCD → GitOps 📊 Kubernetes + Prometheus → Monitoring 📈 Kubernetes + Grafana → Visualization 🧾 Kubernetes + Fluentd → Log Aggregation 🌐 Kubernetes + Istio → Service Mesh 🚦 Kubernetes + NGINX Ingress → Traffic Routing ⚡ Kubernetes + KEDA → Auto & Event-Driven Scaling 🔐 Kubernetes + Vault → Secrets Management 🛡️ Kubernetes + OPA → Policy as Code 💰 Kubernetes + Kubecost → Cost Monitoring 🏢 Kubernetes + Crossplane → Platform Engineering 🔒 Kubernetes + Cilium → Network Security 🤖 Kubernetes + Kubeflow → ML Pipelines 🧪 Kubernetes + MLflow → Experiment Tracking 🚀 Kubernetes + KServe → Model Serving 🧠 Kubernetes + Ollama → LLM Inference 🌍 Kubernetes + Envoy → L7 Traffic Management 🧩 Kubernetes + Inference Gateway → LLM Traffic Routing — Same Kubernetes. Different stacks. Different outcomes. 💥 The truth? Kubernetes didn’t change. 👉 Engineers who understood the ecosystem… won. And now we’re entering the next phase: From DevOps → Platform Engineering → AI Infra If you’re still stuck at “just deploying pods”… You’re already behind. — 💬 Be 𝗵𝗼𝗻𝗲𝘀𝘁: Where are you right now? 1️⃣ Still figuring out Kubernetes 2️⃣ Doing DevOps 3️⃣ Building platforms 4️⃣ Working on AI infra 👇 Comment your stage (no judging) ♻️ Repost if you think Kubernetes is more than just containers #Kubernetes #DevOps #PlatformEngineering #CloudComputing   #AIInfrastructure #Docker #Terraform #GitOps   #SRE #CloudNative #TechLeadership #Engineering #devopinsider 

I’m starting to explore Claude Code and AI coding agents, and this 30-minute workshop triggered an interesting thought for me. Maybe the IDE will become less central in some engineering workflows. If coding agents are moving toward an end-to-end model, the interface may be less about a single editor and more about the full workflow: repository, terminal, tests, Git, documentation, Terraform, CI/CD, logs, containers, and deployment automation. I don’t think IDEs disappear overnight, but for DevOps and Cloud Engineering this CLI-first direction is definitely worth exploring. Original Video: https://lnkd.in/diRrd7zF #ClaudeCode #AIAgents #DevOps #CloudEngineering