<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Willem Berroubache on Medium]]></title>
        <description><![CDATA[Stories by Willem Berroubache on Medium]]></description>
        <link>https://medium.com/@willemberroubache?source=rss-f6dd61eec7d2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*MsiZ0czeMRA-p7XMlX91aQ.jpeg</url>
            <title>Stories by Willem Berroubache on Medium</title>
            <link>https://medium.com/@willemberroubache?source=rss-f6dd61eec7d2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 22:43:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@willemberroubache/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Road to Golden Kubestronaut: Surviving 15 CNCF Certifications]]></title>
            <link>https://medium.com/@willemberroubache/the-road-to-golden-kubestronaut-surviving-16-cncf-certifications-cd8d53524362?source=rss-f6dd61eec7d2------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd8d53524362</guid>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[cncf]]></category>
            <category><![CDATA[certification]]></category>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[cloud-native]]></category>
            <dc:creator><![CDATA[Willem Berroubache]]></dc:creator>
            <pubDate>Wed, 03 Dec 2025 17:30:06 GMT</pubDate>
            <atom:updated>2025-12-03T18:49:28.513Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m6WPw5oJWuum7Xj3y31Vbg.png" /></figure><h3>From Linux fundamentals to Platform Engineering: a deep dive into the ultimate cloud-native challenge.</h3><p>Sixteen Cloud Native certifications. Hundreds of hours of study. Countless late nights staring at a terminal window.</p><p>In <strong>July 2025</strong>, I officially became a <strong>Golden Kubestronaut</strong>, joining a very small handful of engineers worldwide who have completed every single certification in the CNCF track.</p><p>For me, this wasn’t about career advancement. I’ve been working for years across edge computing, 5G security monitoring, AI, cloud platforms, and large-scale orchestration. My background is already deeply rooted in Cloud Security, networking, and Telco expertise.</p><p><strong>Truth be told, I’m long past the stage of needing to prove myself in this domain.</strong> My reputation is established, and my expertise is battle-tested.</p><p>However, I believe that expert status is not a destination, but a foundation. Since the beginning of 2025, I felt a personal need to drastically widen my technical horizon. My goal was not just to “finish the map” of the CNCF, but to reinforce my engineering breadth to tackle the complex architectures of tomorrow.</p><p>This means deepening my capabilities in <strong>Artificial Intelligence, large-scale Service Mesh, and pure software development</strong>, while continuing to explore the edges of the infrastructure landscape — looking beyond standard Kubernetes to solutions like <strong>Nutanix</strong> and hybrid cloud stacks.</p><p>This article is my personal retrospective on going “all the way”: the motivation, the specific hurdles of the hands-on exams, and the failures that forged the success.</p><h3>Why I Decided to Go All the Way</h3><p>As someone deep in Telco and Security, I didn’t <em>need</em> sixteen certifications to do my job. But I wanted to challenge myself to step out of my comfort zone. I wanted to discover the technologies I didn’t use daily to better understand how they integrate with the areas I am focusing on now.</p><p>This path taught me a brutal lesson: <strong>Do not confuse speed with progress.</strong> True expertise only comes from real-world engineering — solving incidents, shipping systems, and failing. The certifications simply validate that you have the vocabulary to do so.</p><h3>The Heavy Hitters: Performance-Based Exams</h3><p>It is crucial to understand that not all CNCF certifications are created equal. Most are multiple-choice, but the core infrastructure exams are <strong>Performance-Based</strong>.</p><p>In these exams, there are no questions to answer A, B, or C. You are given a terminal, a broken cluster (or a blank one), and a list of tasks to solve in a live environment. You either fix the system, or you fail.</p><p>These are the specific requirements for the hands-on labs:</p><p><strong>CKA (Certified Kubernetes Administrator)</strong> The gold standard for cluster operations. It tests troubleshooting, networking, and storage.</p><ul><li><em>Passing Score:</em> <strong>66%</strong></li></ul><p><strong>CKAD (Certified Kubernetes Application Developer)</strong> A race against the clock focusing on deploying, exposing, and debugging applications.</p><ul><li><em>Passing Score:</em> <strong>66%</strong></li></ul><p><strong>CKS (Certified Kubernetes Security Specialist)</strong> The hardest exam of the bunch. It involves deep cluster security, system hardening, and policy enforcement under extreme time pressure.</p><ul><li><em>Passing Score:</em> <strong>67%</strong></li></ul><p><strong>LFCS (Linux Foundation Certified SysAdmin)</strong> The essential groundwork. A practical test of your Linux administration reflexes.</p><ul><li><em>Passing Score:</em> <strong>67%</strong></li></ul><h3>The Knowledge Landscape: The Multiple-Choice Exams</h3><p>The remaining 12 certifications are <strong>Multiple Choice</strong> exams. While less intense than the hands-on labs in terms of pressure, they cover a massive breadth of knowledge across the entire CNCF landscape.</p><p>For almost all of these exams, the standard <strong>Passing Score is 75%</strong>, requiring a solid grasp of the concepts.</p><ul><li><strong>Fundamentals:</strong> The journey starts here with the <strong>KCNA</strong> (Kubernetes and Cloud Native Associate), which lays the conceptual groundwork.</li><li><strong>Security &amp; Policy:</strong> Beyond the CKS, the <strong>KCSA</strong> (Security Associate) covers the security mindset, while <strong>KCA </strong>(Kyverno) focuses on Policy-as-Code and governance.</li><li><strong>Service Mesh &amp; Networking:</strong> This was a key focus for me. <strong>CCA</strong> (Cilium) dives into eBPF networking, while <strong>ICA </strong>(Istio) covers traffic management. <em>Note: The ICA passing score is currently set at </em><strong><em>68%</em></strong><em>.</em></li><li><strong>GitOps &amp; Delivery:</strong> <strong>CGOA</strong> covers GitOps principles, and <strong>CAPA</strong> validates expertise in the Argo ecosystem (CD, Rollouts, Workflows).</li><li><strong>Observability:</strong> <strong>PCA</strong> covers Prometheus/Grafana basics, and <strong>OTCA</strong> deals with OpenTelemetry concepts (more on that one below).</li><li><strong>Platform Engineering:</strong> The newest frontier. <strong>CBA</strong> focuses on Backstage for developer portals, and <strong>CNPA</strong> (passing score <strong>64%</strong>) covers general Platform Engineering strategy.</li></ul><h3>The Reality Check: Overconfidence &amp; The “Slap in the Face”</h3><h4>I have to be honest: it wasn’t a perfect streak.</h4><p>While I sailed through many exams, the <strong>OpenTelemetry (OTCA)</strong> certification was a brutal wake-up call. I failed it on my first attempt. Why? Pure arrogance. I went in too confident, thinking I knew enough about observability to wing it.</p><p>That failure was a slap in the face. It bruised my ego, but it was exactly what I needed. It forced me to stop skimming and really <em>work</em>. I dove back into the subject with twice the intensity, dissecting telemetry pipelines until I truly mastered them. That failure didn’t stop me; it fueled me.</p><p>I also faced a significant challenge with the <strong>ICA (Istio)</strong> exam. I took it before the format change in August 2025. Back then, the 2-hour time limit felt incredibly tight for the depth of questions asked. It was a race against time that tested my composure as much as my technical skills (Passing score was 75% at the time; now adjusted to 68%).</p><h3>Logistics &amp; Longevity: Surviving the PSI Environment</h3><p>Beyond the code, you have to survive the logistics. The exams are proctored via the PSI secure browser, and the rules are strict.</p><p>In the beginning, this was a major source of stress. Clearing the room, removing every scrap of paper, adjusting the webcam, and waiting for the proctor can feel daunting. But after sixteen exams, I can tell you: <strong>it becomes a routine.</strong></p><p>My advice is to prepare your environment well in advance. Create a minimalist “exam zone.” Once you eliminate the environmental variables, the pre-exam check-in becomes just another step in the process, allowing you to focus entirely on the terminal.</p><p>It is also worth noting that this journey doesn’t have a finish line. Most of these certifications expire after two or three years. Becoming a Kubestronaut is not a one-time achievement; it is a commitment to a continuous cycle of learning and renewal.</p><h3>How I Prepared (My Stack)</h3><p>People often ask for the “magic resource.” The truth is, there is no secret shortcut. My method was consistent:</p><ol><li><strong>Udemy:</strong> This was my primary driver for structured learning.</li><li><strong>Documentation:</strong> I forced myself to read the official documentation for every solution. Courses give you the summary; docs give you the “why.”</li><li><strong>Relentless Labs:</strong> I didn’t just watch videos. I spent hundreds of hours in terminal windows (<strong>the real life!</strong>). If I couldn’t build it without looking at the guide, I wasn’t ready.</li></ol><h3>The New Golden Benefits</h3><p>The “Golden Kubestronaut” status carries prestige, and the CNCF recently overhauled the benefits program. As of late 2025, Golden Kubestronauts receive exclusive gear, a ThriveOne wellness subscription, free Kubernetes Community Days (KCD) tickets, and heavy discounts on KubeCon and future certifications.</p><p>Even better, starting in <strong>2026</strong>, the program will include a professional recommendation letter, turning this from a technical achievement into a serious career asset.</p><h3>Final Thoughts: A Transformation</h3><p>Becoming a Golden Kubestronaut was a personal transformation. It taught me to think holistically about platforms, automation, and developer experience. But now that I have reached the summit of Cloud Native, I am turning my gaze toward the next challenges.</p><p>The reality of large enterprises is rarely “pure public cloud.” It is messy, hybrid, and demanding. That is why I am now digging deeper into solutions like Nutanix Kubernetes Platform (NKP) and on-premise architectures, bridging the gap between traditional infrastructure, AI Agents, and complex Mesh networking.</p><p>The future belongs to those who can connect these worlds. That is exactly where I intend to be.</p><p><strong>The cloud-native world evolves fast, but I intend to stay ahead of the curve, not by collecting badges, but by building systems that matter.</strong></p><p>I’ll see you in orbit. 🚀</p><p><strong>Proof of the journey:</strong> You can view my full certification portfolio here: <a href="https://www.credly.com/users/willem">Credly Profile</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd8d53524362" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ HOLD UP—AI CAN WORK NOW?!? Say Hello to MCP.]]></title>
            <link>https://medium.com/@willemberroubache/ai-that-does-stuff-say-hello-to-mcp-dad522b0d575?source=rss-f6dd61eec7d2------2</link>
            <guid isPermaLink="false">https://medium.com/p/dad522b0d575</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[orchestration]]></category>
            <category><![CDATA[mcps]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[model-context-protocol]]></category>
            <dc:creator><![CDATA[Willem Berroubache]]></dc:creator>
            <pubDate>Sat, 29 Mar 2025 01:23:55 GMT</pubDate>
            <atom:updated>2025-03-29T11:23:44.477Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>🧠 HOLD UP – AI CAN WORK NOW?!? Say Hello to MCP.</strong></h3><p>You’ve got a powerful AI model — it can reason, explain, even write poetry if you ask nicely. But when it comes to actually <strong>doing things</strong> in the real world — like sending a message, triggering a deployment, or querying a live database — things get complicated fast.</p><p>That’s where the <strong>Model Context Protocol (MCP)</strong> comes in. Think of it as the missing link between smart models and real tools. It’s not just another integration layer — it’s a lightweight, modular protocol that gives your AI <strong>context, control, and capabilities</strong>.</p><p>In this article, we’ll unpack what MCP is, how it works, and why it might just be the most important thing you haven’t tried yet in the world of AI deployment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BrMv-iubeVq-w0kcuSPiMQ.png" /></figure><h3><strong>1. What is MCP?</strong></h3><p>In slightly more technical terms, MCP is like an API (Application Programming Interface) for AI models. But way cooler. It’s a standardized way to talk to your AI, telling it what to do, when to do it, and how to report back.</p><p>In an ecosystem where Large Language Models (LLMs) are becoming increasingly essential, connecting these AIs to a variety of tools and data sources is a necessity. This is where the <strong>Model Context Protocol</strong> comes in.</p><p>The Model Context Protocol is an open-source project run by Anthropic, PBC, and open to contributions from the entire community. Anthropic is a Public Benefit Corporation (PBC) dedicated to building reliable, interpretable, and steerable AI systems. They are best known for developing Claude, a powerful LLM.</p><p>The Model Context Protocol is designed to facilitate the transmission of context between applications and language models. Its goal is to enable an LLM to easily connect to different data sources and tools, much like connecting various peripherals to a computer via a single type of universal port. This simplifies the ecosystem, reduces development costs, and gives technical teams the freedom to choose their model providers or tools.</p><h4><strong>The Model Context Protocol: Connecting LLMs to the World</strong></h4><p>Connecting LLMs to the data and tools they need is often a real headache: specific APIs, in-house solutions, lack of clear standards, etc. It’s to address these challenges that the Model Context Protocol (MCP) was created.</p><p>In this section, we’ll explore how MCP facilitates this integration by providing a standardized and flexible framework, allowing LLMs to easily connect to local and remote data sources. We’ll explore its architecture, its benefits, and its potential to simplify the management of complex workflows, while making AI agents more autonomous and performant.</p><h3><strong>2. MCP Architecture: A Client-Server Symphony</strong></h3><p>At the heart of MCP lies a client-server architecture. MCP hosts (e.g., a development platform, a code editing environment, or a business tool) connect via standardized clients and communicate with MCP servers, each specializing in a specific type of integration or capability. You can have a server dedicated to connecting to your internal database, another for interacting with an external service via an API, or even a third for accessing your local files.</p><p>The whole thing forms a central “hub,” where each server exposes its functionality in a standardized way. The AI agent or LLM used can then easily switch between different sources, tools, and services without having to rewrite integration code for each new need.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1012/1*6tM-1JKBYqr25NKlPOcRuA.png" /></figure><p>As the diagram illustrates, MCP follows a client-server architecture where a host application can connect to multiple servers:</p><ul><li><strong>MCP Hosts:</strong> Programs like IDEs, AI tools, or Claude, that want to access data through MCP.</li><li><strong>MCP Clients:</strong> Protocol clients that maintain 1:1 connections with servers.</li><li><strong>MCP Servers:</strong> Lightweight programs that each expose specific capabilities through the standardized Model Context Protocol.</li><li><strong>Local Data Sources:</strong> Your computer’s files, databases, and services that MCP servers can securely access.</li><li><strong>Remote Services:</strong> External systems available over the internet (e.g., through APIs) that MCP servers can connect to.</li></ul><h3><strong>3. Okay, We Get It, But How Does It <em>Really</em> Work?</strong></h3><p>Instead of just one integration, let’s take things to the next level and demonstrate the true power of MCP by connecting Claude to three different services simultaneously.</p><ul><li><a href="https://github.com/Flux159/mcp-server-kubernetes"><strong>Kubernetes</strong></a><strong> :</strong> To deploy some workloads.</li><li><a href="https://github.com/varunneal/spotify-mcp"><strong>Spotify</strong></a><strong>:</strong> To control music playback.</li><li><a href="https://github.com/modelcontextprotocol/servers/tree/main/src/slack"><strong>Slack</strong></a><strong>:</strong> To send notifications and interact with our team.</li></ul><p>This example will show how MCP can act as a central hub, allowing Claude to orchestrate complex workflows across diverse technologies.</p><p><strong>You don’t need a paid subscription for this demo! The free version of Claude will work just fine.</strong></p><p><strong>Disclaimer:</strong> This is a simplified example for demonstration purposes. In a real-world scenario, you would need to handle authentication, security, and error handling more robustly. Also, you’ll need accounts and API keys for Spotify and Slack.</p><p><strong>First Steps: Getting Claude Ready (MacOS)</strong></p><p>Before we dive into the technical details, you’ll need to get Claude set up on your MacOS machine:</p><ol><li><strong>Download and Install Claude:</strong> Head over to <a href="https://claude.ai/">https://claude.ai/</a> and download the Claude application for MacOS.</li><li><strong>Sign Up or Log In:</strong> Once installed, launch Claude and either sign up for a new account or log in with your existing credentials.</li><li><strong>Configure MCP:</strong> This is where we tell Claude to listen for our custom MCP servers. Open Claude and navigate to <strong>Settings -&gt; Developer</strong>. This is where you’ll add the connection details for each of our servers later.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x2FQ3b-Jgk9UQhUuHD2M2w.png" /></figure><p>Then you can click on <strong>“Edit Config”</strong>. This will open a window containing the <strong>“claude_desktop_config.json”</strong> file. This file holds the configuration that allows you to choose which MCP servers to use. Edit it with your favorite text editor.</p><h4><a href="https://github.com/Flux159/mcp-server-kubernetes">Let’s Add Kubernetes MCP Server</a>:</h4><p>https://github.com/Flux159/mcp-server-kubernetes</p><p>Before diving into the server setup, let’s ensure you have <strong>the necessary tools and access</strong>. Here’s what you’ll need:</p><ul><li>A Kubernetes cluster (obviously…)</li><li>kubectl installed and in your PATH</li><li>A valid kubeconfig file with contexts configured</li><li>Access to a Kubernetes cluster configured for kubectl (e.g., minikube, Rancher Desktop, GKE, etc.)</li><li>Helm v3 installed and in your PATH (no Tiller required). Optional if you don’t plan to use Helm.</li></ul><p>As you can see, everything seems OK! My cluster and kubectl are ready!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SUfMUN1H8GYkqRGm1QBRvg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_Z4SAKQYwuL4-pA74yyIWQ.png" /></figure><p><strong>Now we can add our Kubernetes MCP server to our Claude client</strong>. It’s simple! We’ll revisit our “<strong>claude_desktop_config.json”</strong> file and add the following code:</p><pre>{<br>    &quot;mcpServers&quot;: {<br>      &quot;kubernetes&quot;: {<br>        &quot;command&quot;: &quot;npx&quot;,<br>        &quot;args&quot;: [&quot;mcp-server-kubernetes&quot;]<br>      }<br>    }<br>}</pre><p>Next, <strong>restart your Claude client</strong> and return to the developer settings. The MCP server should now be in “Ready” status, as shown in the screenshot below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qAjJN2qP4evk7_GoT8O9DQ.png" /></figure><p>Now, return to your conversation window. <strong>You should see a new icon indicating that MCP is now available</strong>, along with its capabilities. We can now query our cluster based on the MCP server’s capabilities. Let’s ask for a list of pods in the cluster and witness the magic!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KzDflAoa1einR8tpnCD_aw.png" /></figure><p>Claude will also ask you to authorize the use of MCP for the current request or the entire chat session.</p><p>After allow MCP, you can admire the response:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mBJzkIx3_nBfRav0vUd13g.png" /></figure><h3><strong>Now, let’s use a workflow that interacts with other MCP servers. We’ll use the following scenario:</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eOLhIa9dGxw7O-3AUal19w.png" /></figure><p>This prompt automates the deployment of a web server in the web-front namespace, sends pod logs and deployment updates to Slack, and celebrates success with 2Pac’s California Love on Spotify. It’s a fun mix of DevOps, notifications, and good vibes.</p><blockquote>Deploy a web server in the existing web-front namespace.</blockquote><blockquote>Once deployed and running, send pods logs to a Slack channel to ensure the server is ready and notify the team through slack.</blockquote><blockquote>Send deployment and pods created to the team.</blockquote><blockquote>Once everything is done, it’s party time, start 2Pac — California Love on Spotify!</blockquote><p><strong>Before launching it, you need to configure the three MCPs.</strong></p><p>Just grab the documentation for the <a href="https://github.com/modelcontextprotocol/servers/tree/main/src/slack">Slack</a> and <a href="https://github.com/varunneal/spotify-mcp">Spotify</a> MCPs and enter your API keys.</p><p>For Spotify, make sure to configure in the Spotify Developer Portal the callback URL to point to your localhost and the port defined in your configuration.</p><p>The claude_desktop_config.json file should look like this:</p><pre>{<br>  &quot;mcpServers&quot;: {<br>    &quot;kubernetes&quot;: {<br>      &quot;command&quot;: &quot;npx&quot;,<br>      &quot;args&quot;: [&quot;mcp-server-kubernetes&quot;]<br>    },<br>    &quot;slack&quot;: {<br>      &quot;command&quot;: &quot;npx&quot;,<br>      &quot;args&quot;: [<br>        &quot;-y&quot;,<br>        &quot;@modelcontextprotocol/server-slack&quot;<br>      ],<br>      &quot;env&quot;: {<br>        &quot;SLACK_BOT_TOKEN&quot;: &quot;xoxb-XXXXXXXXX-XXXXXXXX-XXXXXX&quot;,<br>        &quot;SLACK_TEAM_ID&quot;: &quot;T0XXXXXXXXXXXXX&quot;<br>      }<br>    },<br>    &quot;spotify&quot;: {<br>      &quot;command&quot;: &quot;uv&quot;,<br>      &quot;args&quot;: [<br>        &quot;--directory&quot;,<br>        &quot;/Users/me/workspace/mcp/spotify-mcp/src&quot;,<br>        &quot;run&quot;,<br>        &quot;spotify-mcp&quot;<br>      ],<br>      &quot;env&quot;: {<br>        &quot;SPOTIFY_CLIENT_ID&quot;: &quot;XXXXXXXXXXXXX&quot;,<br>        &quot;SPOTIFY_CLIENT_SECRET&quot;: &quot;XXXXXXXXXXX&quot;,<br>        &quot;SPOTIFY_REDIRECT_URI&quot;: &quot;http://localhost:8889&quot;<br>      }<br>    }<br>  }<br>}</pre><p>As shown in the screenshot, all three are OK.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jQA-GxyokhKFiSZ_vzYYPQ.png" /></figure><h3><strong>TADAA!</strong> 🎉 You’re now the king of MCP and prompting!</h3><p>Everything should run smoothly from here.</p><p>Claude will ask for permissions to use the different MCP servers — just make sure to accept them, obviously. 👑⚙️</p><p>The video below is a full demo of the setup (feel free to nod your head to the music — no one’s judging 😎🎶).</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F6Y0xHBrpTPI%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D6Y0xHBrpTPI&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F6Y0xHBrpTPI%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/8755a57c0017afef68b8bc38bde9a28b/href">https://medium.com/media/8755a57c0017afef68b8bc38bde9a28b/href</a></iframe><h3><strong>4. Explore the MCP Universe: There’s a Server for Everything!</strong></h3><p>The world of MCP is rich, modular, and constantly evolving. Whether you’re working with files, databases, code, or cloud platforms, there’s likely an MCP server ready to plug into your workflow. Official reference implementations cover essentials like <strong>PostgreSQL</strong>, <strong>Google Drive</strong>, <strong>SQLite</strong>, and <strong>GitHub</strong>. Development tools like <strong>Git</strong>, <strong>Sentry</strong>, and <strong>GitLab</strong> make project management seamless. Want to automate browsing or search the web? Servers like <strong>Brave Search</strong>, <strong>Fetch</strong>, and <strong>Puppeteer</strong> have you covered. On the productivity side, tools like <strong>Slack</strong>, <strong>Todoist</strong>, and <strong>Google Maps</strong> bring communication and task tracking into the loop. And let’s not forget fun or specialized integrations like <strong>Spotify</strong>, <strong>Sequential Thinking</strong>, and <strong>EverArt</strong>. With official platform support from names like <strong>Cloudflare</strong>, <strong>Stripe</strong>, <strong>Neon</strong>, and more, plus an active community building new servers for tools like <strong>Docker</strong>, <strong>Kubernetes</strong>, and <strong>Snowflake</strong>, MCP offers a powerful ecosystem where LLMs can truly act.</p><h3><strong>5. Claude… and beyond! 🚀</strong></h3><p>In this example, we’re using <strong>Claude</strong> as the LLM host, but the beauty of MCP is its flexibility — you can use it with many other compatible hosts that support the protocol. The logic stays the same: the host acts as the brain, and the MCP servers act as the hands, eyes, and ears. In just a few days, I’ll share another, more concrete example to dive deeper into real-world use cases and show how this setup can supercharge your workflows, no matter which LLM you prefer.</p><h3><strong>6. Okay, but what’s next, man?</strong></h3><p>So you’ve seen MCP in action — it’s slick, powerful, and already useful. But this is just the warm-up. The real magic? It’s what’s coming next. MCP isn’t just a protocol for plugging LLMs into tools — it’s a <strong>blueprint for building truly interactive, context-aware AI systems</strong> that can live <em>anywhere</em> and talk to <em>anything</em>.</p><p><strong>📡 Private mobile networks, meet your new AI operator</strong></p><p>One of the most exciting frontiers? <strong>Private mobile networks</strong>. Think industrial campuses, factories, or logistics hubs running on private 5G — MCP could be the bridge that lets AI agents directly interact with operational tech: machines, logs, sensors, dashboards. Instead of passive monitoring, your AI could <em>query</em>, <em>control</em>, and <em>orchestrate</em> processes across the network, in real time, with fine-grained access control. No middle layers. No hacky workarounds. Just clean, protocol-based action.</p><p><strong>🏢 Bring Your Own Infrastructure (and Let the AI Do the Rest)</strong></p><p>Enterprises spend fortunes integrating systems. MCP flips the script. With standard interfaces to tools like GitHub, Slack, Postgres, and even Sentry or Stripe, you can let AI agents operate directly in your stack — whether it’s cloud-native, hybrid, or running in a basement somewhere. Want a bot that monitors incidents, summarizes them, creates Jira tickets, and sends updates on Slack? Easy. And it’s all traceable, permissioned, and host-agnostic.</p><p><strong>🧠 LLMs with actual hands, not just vibes</strong></p><p>Most LLMs are still stuck in the “suggestion box” phase — they can talk, but they can’t <em>do</em>. MCP gives them tools. Imagine agents that don’t just generate ideas, but actually <strong>run the build, write the migration script, push the commit, update the doc, and ping your team</strong>. All from a local, secure, auditable setup. This isn’t AI-as-a-service — this is <strong>AI-as-a-coworker</strong>.</p><p><strong>🧩 One protocol to plug them all</strong></p><p>Here’s the bigger picture: MCP isn’t locked to any tool, platform, or provider. It’s a protocol. That means if you build a new tool — for IoT, finance, manufacturing, space missions, whatever — you can expose it via an MCP server and any compatible LLM can immediately start using it. This opens up a path to <strong>true interoperability</strong> between AI and the real world. It’s like giving your AI a universal adapter for everything.</p><p>MCP is still early, but it’s quietly laying the foundation for a future where AI isn’t boxed into chats and demos. It’s <strong>in the loop</strong> — reading, acting, reacting, and improving systems around it. If you’re building infrastructure, deploying private networks, or automating complex workflows, the question isn’t <em>if</em> you’ll use MCP… it’s <em>how soon</em>.</p><h3><strong>🚀 We’re Just Getting Started</strong></h3><p>Now that you’ve got a taste of MCP and how it unlocks real-world actions for AI models, buckle up — there’s much more coming. In future articles, I’ll dive deeper into concrete examples and more advanced use cases, tackling increasingly complex scenarios.</p><p>Expect hands-on deployments, sophisticated workflows, and practical tips that’ll elevate your prompting game. If you’re ready to level-up your AI toolkit, stay tuned — this was just the warm-up!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dad522b0d575" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Road to Kubestronaut Status: How We Rocketed Through the CNCF Challenge]]></title>
            <link>https://medium.com/the-road-to-kubestronaut-status-how-we-rocketed/the-road-to-kubestronaut-status-how-we-rocketed-through-the-cncf-challenge-c33275330486?source=rss-f6dd61eec7d2------2</link>
            <guid isPermaLink="false">https://medium.com/p/c33275330486</guid>
            <category><![CDATA[kubestronaut]]></category>
            <category><![CDATA[certification]]></category>
            <category><![CDATA[cncf]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[cloud-native]]></category>
            <dc:creator><![CDATA[Willem Berroubache]]></dc:creator>
            <pubDate>Mon, 23 Sep 2024 19:52:29 GMT</pubDate>
            <atom:updated>2024-09-23T21:36:21.905Z</atom:updated>
            <content:encoded><![CDATA[<p>Story co-written with <a href="https://medium.com/u/921e34038048">Frederic Klein</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*75CI642qnW_sfGI-DPJgvQ.jpeg" /></figure><h3>Introduction</h3><p>Achieving the <a href="https://www.cncf.io/training/kubestronaut/">Kubestronaut</a> status, awarded by the Cloud Native Computing Foundation (CNCF), is quite challenging. This title is for people who have shown expertise in Kubernetes and cloud-native technologies. Our journey began as a personal challenge between myself, and my friend, <a href="https://medium.com/u/921e34038048">Frederic Klein</a>.</p><p>For me, it was an opportunity to validate my technical skills and gain recognition from my employer.</p><p>In this article, we share our journey, the challenges we faced, and the resources that helped us reach this important milestone in the world of Kubernetes and cloud-native technologies</p><p><strong>What is </strong><a href="https://www.cncf.io/training/kubestronaut/"><strong>Kubestronaut</strong></a><strong> Status?</strong></p><p>To become a <a href="https://www.cncf.io/training/kubestronaut/">Kubestronaut</a>, candidates must simultaneously hold the following five certifications, each valid for two years:</p><ol><li><a href="https://training.linuxfoundation.org/certification/certified-kubernetes-administrator-cka/"><strong>Certified Kubernetes Administrator (CKA)</strong></a><strong>:</strong> Validates expertise in managing and administering Kubernetes clusters.</li><li><a href="https://training.linuxfoundation.org/certification/certified-kubernetes-application-developer-ckad/"><strong>Certified Kubernetes Application Developer (CKAD)</strong></a><strong>:</strong> Demonstrates skills in designing, building, and deploying applications on Kubernetes.</li><li><a href="https://training.linuxfoundation.org/certification/certified-kubernetes-security-specialist/"><strong>Certified Kubernetes Security Specialist (CKS)</strong></a><strong>:</strong> Confirms knowledge in securing container-based applications and Kubernetes environments.</li><li><a href="https://training.linuxfoundation.org/certification/kubernetes-cloud-native-associate/"><strong>Kubernetes Cloud Native Associate (KCNA)</strong></a><strong>:</strong> Covers essential knowledge of cloud-native technologies, focusing on Kubernetes for newcomers.</li><li><a href="https://training.linuxfoundation.org/certification/kubernetes-and-cloud-native-security-associate-kcsa/"><strong>Certified Kubernetes Security Associate (KCSA)</strong></a><strong>:</strong> Shows understanding of Kubernetes security and how to enhance cluster security.</li></ol><p>Each certification exam is conducted remotely. The KCNA and KCSA are multiple-choice exams, while the CKA, CKAD, and CKS are practical, hands-on tests performed on a remote Linux VM connected to Kubernetes clusters.</p><p><strong>Our Difficulty Ranking of the Exams</strong></p><p>From easiest to hardest, we ranked the exams as follows:</p><ol><li><strong>KCNA</strong></li><li><strong>CKAD</strong></li><li><strong>CKA</strong></li><li><strong>KCSA</strong></li><li><strong>CKS</strong></li></ol><p>The foundational concepts in the KCNA made it the least challenging, while the advanced security topics in the KCSA and CKS required a deeper, more comprehensive understanding.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/802/1*wMLPJ22Al-2F-cG8l5gluw.png" /></figure><p><strong>Effective Preparation Strategies</strong></p><ol><li><strong>Study Official Documentation:</strong></li></ol><ul><li><strong>Kubernetes Documentation:</strong> Dive deep into the official docs to understand core concepts, architecture, and best practices.<a href="https://kubernetes.io/docs"> Kubernetes Docs</a></li><li><strong>Kubernetes Blog:</strong> Stay updated with the latest developments and insights.<a href="https://kubernetes.io/blog/"> Kubernetes Blog</a></li></ul><p>2. <strong>For the CKS exam, additional allowed resources include:</strong></p><ul><li><a href="https://aquasecurity.github.io/trivy/v0.55/">Trivy Documentation</a></li><li><a href="https://falco.org/docs/">Falco Documentation</a></li><li><a href="https://etcd.io/docs/">etcd Documentation</a></li><li><a href="https://wiki.ubuntu.com/AppArmor">AppArmor Documentation</a></li><li><a href="https://helm.sh/docs/">Helm Documentation</a> (also allowed in CKAD)</li></ul><p><strong>3. Gain Hands-On Experience:</strong></p><ul><li><strong>Practical Usage:</strong> Regularly use Kubernetes in real-world scenarios to solidify your understanding.</li><li><strong>Experimentation:</strong> Set up your own clusters and explore different features and configurations.</li></ul><p><strong>4. Self-Care is Essential:</strong></p><ul><li><strong>Manage Stress:</strong> Take regular breaks and maintain a healthy study-life balance.</li><li><strong>Take Your Time:</strong> Don’t rush your preparation. Thorough understanding trumps speed.</li><li><strong>Exam Day Prep:</strong> Ensure you’re well-rested and familiar with the exam procedures.</li></ul><h3><strong>Tips and Tricks for Success</strong></h3><p><em>For Multiple-Choice Exams (KCSA and KCNA):</em></p><ul><li><strong>Understand the Exam Format:</strong> Familiarize yourself with question types and structure.</li><li><strong>Focus on Key Concepts:</strong> Grasp the core principles of Kubernetes and cloud-native technologies.</li><li><strong>Read Carefully:</strong> Thoroughly read each question and all possible answers.</li><li><strong>Manage Your Time:</strong> Use the flagging feature for difficult questions and revisit them later.</li></ul><p><em>For Practical Exams (CKA, CKAD, and CKS):</em></p><ul><li><strong>Set Up a Practice Environment:</strong> Use tools like Minikube or Kind for local clusters.</li><li><strong>Structured Study Plan:</strong> Break down the syllabus and allocate time for each topic.</li><li><strong>Practice with Killer.sh:</strong> Utilize the simulator for realistic exam scenarios. Two sessions are included with your exam purchase.</li><li><strong>Quick Navigation of Documentation:</strong> Practice finding information swiftly in the allowed resources.</li><li><strong>Simulate Exam Conditions:</strong> Practice under timed conditions to build confidence.</li></ul><h3>The most important points</h3><ol><li><strong>Verify Your Prompt:</strong> Always ensure you’re working in the correct environment.</li><li><strong>Use Command Aliases:</strong> Create aliases (e.g., alias k=kubectl) to speed up command execution.</li><li><strong>Organize Terminal Tabs:</strong> Use separate tabs for different tasks to avoid confusion.</li><li><strong>Efficient Copy-Paste:</strong> Master keyboard shortcuts for copying and pasting within the terminal and browser.</li><li><strong>Leverage Imperative Commands:</strong> Use imperative commands to reduce YAML editing time.</li><li><strong>Utilize Command History:</strong> Use CTRL+R to search your command history efficiently.</li><li><strong>Search Functions:</strong> Use browser search features to quickly find information in documentation.</li><li><strong>Don’t be afraid to click on unauthorized links</strong>: If you click on a forbidden link, don’t panic; you will receive an error message indicating that the resource is not allowed, but you will not be affected and the exam will continue.</li><li><strong>Know Your Editor:</strong> Familiarize yourself with vi editor commands to edit files swiftly.</li></ol><h4>Always verify your prompt</h4><p>For some questions, you will need to write your responses in files. These responses <strong>MUST ALWAYS BE WRITTEN ON THE REMOTE DESKTOP VM</strong>. If your prompt indicates that you are on a cluster, this is incorrect, and you will not receive points. If you have a prompt like this one, you’re <strong>WRONG</strong> :</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/0*zMToaMi_NIVO9XY2" /><figcaption>Wrong prompt for write responses</figcaption></figure><p>Always ensure your prompt is showing the correct environment before starting your task.</p><p>The answers to be written on the remote machine should be saved in the following path: /opt/course/&lt;Question Number&gt;</p><h4>Utilize Keyboard Shortcuts and Command Aliases</h4><p>Use keyboard shortcuts and command aliases to save time during the exam. For instance, create an alias for kubectl by adding alias k=kubectl to your shell configuration. This allows you to execute commands more quickly, such as k get pods instead of kubectl get pods or k describe pod &lt;pod_name&gt; instead of kubectl describe pod &lt;pod_name&gt;. For applying configurations, you can use k apply -f &lt;file.yml&gt;. It’s also helpful to create shortcuts for common resource edits with commands like k edit &lt;resource&gt;/&lt;name&gt;. Additionally, use tab completion to quickly fill in commands and file names.</p><p>If you’re working across multiple namespaces, switch between them efficiently using k config set-context — current — namespace=&lt;namespace&gt; or view all resources across all namespaces with k get pods -A. These shortcuts can significantly speed up your command-line interactions, helping you focus on solving more complex problems during the exam.</p><h4>Terminal Tabs for Different Tasks</h4><p>For terminal usage, it’s beneficial to open two terminal tabs. Use one tab specifically for querying clusters via kubectl, and use the second tab solely for SSH connections. This separation helps keep your workspace organized and allows you to quickly switch between tasks without confusion.</p><h4>Utilize a Notepad on the Remote Machine</h4><p>Having a notepad open on the remote machine can be incredibly useful for copying and pasting commands, configurations, and notes. This can help you avoid mistakes and streamline your workflow.</p><h4>Always Copy Paste Cluster and Context Changes</h4><p>When switching between questions that involve different clusters or contexts, always copy and paste the commands for changing clusters and contexts to avoid any mistakes. This ensures you are working within the correct environment and reduces the risk of executing commands in the wrong cluster. For example, use commands like kubectl config use-context &lt;context_name&gt; and kubectl config set-context — current — namespace=&lt;namespace&gt; to ensure you are in the right context and namespace for each question.</p><h4>Efficient Copy and Paste</h4><p>When it comes to copying and pasting, here are some essential tips to keep in mind:</p><p><strong>Copying from the Question Panel:</strong></p><ul><li>When a <strong>Copy</strong> link is present, click it with the mouse to copy the text.</li><li>If the text is highlighted, clicking on it will automatically copy it.</li><li>Note that hovering over the text may show the <strong>Copy</strong> link, but you can’t actually click the link itself.</li></ul><p><strong>Copy and Paste Shortcuts:</strong></p><ul><li><strong>Terminal Emulator:</strong> Use CTRL + SHIFT + C to copy and CTRL + SHIFT + V to paste.</li><li><strong>Firefox:</strong> Use CTRL + C to copy and CTRL + V to paste.</li><li><strong>Other:</strong> Right-click to bring up the context menu and select copy or paste.</li></ul><p><strong>Text Selection in Terminal:</strong></p><ul><li><strong>Shift + Double Click:</strong> Select words, file paths, etc.</li><li><strong>Shift + Drag:</strong> Select multiple words or lines.</li></ul><h4>Master Imperative Commands</h4><p>Knowing imperative commands can greatly reduce the amount of YAML editing you need to do, saving you valuable time. For instance, if you are asked to create an httpd pod with a volume mount, create the pod using imperative commands first and then edit the YAML for any additional configurations.</p><p>Here’s an example workflow:</p><ol><li>Create the pod imperatively and output the YAML:</li></ol><p>k run my-nginx — image=httpd — dry-run=client -o yaml &gt; my-pod.yaml</p><p>2. Open the YAML file for editing:</p><p>vi my-pod.yaml</p><p>3. Open the YAML file for editing:</p><p>k apply -f my-pod.yaml</p><h4>Utilize Terminal Command History</h4><p>Instead of repeatedly pressing the up-arrow to scroll through your command history, use the command history search feature with CTRL+R to quickly locate the information you need. This can be especially helpful when you need to reference specific commands, configurations, or explanations during the exam. Make sure to familiarize yourself with advanced search techniques to maximize the benefits of this feature.</p><h4>Master vi Editor Basics</h4><p>Knowing how to efficiently use the vi editor can save you significant time. Key commands include:</p><ul><li><strong>Visual Line (SHIFT+V)</strong>: Selects a block of text.</li><li><strong>y (yank)</strong>: Copies the selected text.</li><li><strong>d (delete)</strong>: Cuts the selected text.</li><li><strong>p (paste)</strong>: Pastes the copied or cut text.</li></ul><p>Remember, these commands operate within vi’s own paste buffer, not the system clipboard.</p><h3>Useful Links and Resources</h3><h4>Training and learning</h4><ul><li><a href="https://kodekloud.com/courses">KodeCloud</a></li><li><a href="https://www.udemy.com/courses/search/?src=ukw&amp;q=kubernetes+certification">Udemy</a></li><li><a href="https://killer.sh/">killer.sh</a></li><li><a href="https://killercoda.com/killer-shell-cka">Study all scenarios on Killercoda</a></li></ul><h4>CKA/CKAD</h4><ul><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/tips-cka-and-ckad#cka-and-ckad-environment">Exam Environment</a></li><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/certification-resources-allowed#certified-kubernetes-administrator-cka-and-certified-kubernetes-application-developer-ckad">Allowed Documentation</a></li></ul><h4>CKS</h4><ul><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/important-instructions-cks#cks-environment">Exam Environment</a></li><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/certification-resources-allowed#certified-kubernetes-security-specialist-cks">Allowed Documentation</a></li></ul><h4>KCNA / KCSA</h4><ul><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/frequently-asked-questions-kcna">Exam Environment</a></li><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/important-instructions-kcna">Instructions</a></li></ul><h4>Before Exam</h4><ul><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/faq-cka-ckad-cks#what-are-the-system-requirements-to-take-the-exam">Exam System Requirements</a></li><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/faq-cka-ckad-cks#what-are-the-testing-environment-requirements-to-take-the-exam">Exam Workspace Requirements</a></li><li><a href="https://docs.linuxfoundation.org/tc-docs/certification/faq-cka-ckad-cks#what-are-the-id-requirements-to-take-the-exam">Exam ID requirements</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c33275330486" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-road-to-kubestronaut-status-how-we-rocketed/the-road-to-kubestronaut-status-how-we-rocketed-through-the-cncf-challenge-c33275330486">The Road to Kubestronaut Status: How We Rocketed Through the CNCF Challenge</a> was originally published in <a href="https://medium.com/the-road-to-kubestronaut-status-how-we-rocketed">The Road to Kubestronaut Status: How We Rocketed Through the CNCF Challenge</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>