GitHub Copilot Introduces Usage Limits for Reliability

[New blog post] Packaging GitHub Copilot Agents and Skills with Agent Package Manager If you are building GitHub Copilot agents and skills across multiple repositories, creating them is usually the easy part. Keeping them consistent across teams and repos is where things start to get messy. 📦 In this post, I look at how Agent Package Manager gives you a cleaner way to package and share GitHub Copilot agents, skills, and supporting MCP configuration without relying on copy-and-paste across .github folders. ⚡ 🔗 https://lnkd.in/ec_uE9jJ #GitHubCopilot #GitHub #AgentPackageManager #MCP #PlatformEngineering #DevOps #DeveloperExperience #AIEngineering

[New Blog Post] What Makes a Good GitHub Copilot Agent Skill? In this post, I break down what I think makes a good GitHub Copilot agent skill, including: - giving the agent the right context - shaping behaviour, not just adding information - keeping skills focused and predictable - avoiding over-engineered instructions - designing for real engineering use, not demo scenarios If you are starting to build skills, or refining the ones you already have, this may help you think a bit more clearly about what actually makes them effective in practice. Read here: https://lnkd.in/ezxctYxE #GitHubCopilot #GitHub #AgentSkills #AIEngineering #PlatformEngineering #DeveloperExperience #DevOps #CloudEngineering

The first time I pushed to GitHub after implementing GitOps in my home lab, I just sat there watching my terminal. I was not sure it was going to work. Then it did. The change applied itself. No kubectl apply. No manual anything. Flux just picked it up and made it happen. I had read about GitOps. I had watched videos about GitOps. But seeing your cluster update itself from a Git push is a completely different feeling. Here is what happened under the hood: - I pushed an updated manifest to GitHub. - Flux detected the change within seconds. - Flux compared what was in Git to what was running in the cluster. - A new pod spun up with the updated config. The old one terminated cleanly. I did not touch the cluster once. That was the moment I understood why teams and enterprises adopt GitOps as an industry standard. The cluster can only ever be in a state that exists in Git. No undocumented changes. No mystery configs. Every change has a commit. Every commit has a history. Every history has a reason. Once you work this way it is very hard to go back to doing things manually. Have you had a moment where a tool completely changed how you think about your workflow? 👇 Follow me, I am documenting everything I build and learn in my home lab. #GitOps #Kubernetes #DevOps #FluxCD #CloudNative

Question for all the engineers out there; Is GitHub still the undisputed king of Git platforms? 👑💻 In a recent deep-dive by The Pragmatic Engineer, a fascinating question was raised: Does GitHub still merit the title of the top Git platform? For years, GitHub has been the default choice for developers, open-source contributors, and enterprise engineering teams worldwide. But as the developer tooling landscape shifts, that "default" status is being challenged. Here are a few key factors making engineering leaders pause and re-evaluate their tech stacks: 1️⃣ The Rise of Robust Alternatives: Platforms like GitLab have drastically evolved, offering incredibly powerful, all-in-one CI/CD pipelines and DevOps workflows that aggressively rival GitHub Actions. 2️⃣ Reliability & Uptime: With massive scale comes intense scrutiny. Occasional outages are a stark reminder to engineering teams about the risks of single points of failure. 3️⃣ The AI Factor: GitHub Copilot undeniably revolutionized developer productivity, but competitors are rapidly launching their own AI-driven assistants to close the gap. 4️⃣ Enterprise Costs: As engineering orgs scale and tighten their belts, pricing models are being heavily scrutinized, prompting some teams to explore self-hosted or more cost-effective alternatives. The takeaway? GitHub remains an absolute powerhouse, but choosing it is no longer just a "no-brainer." Modern engineering teams need to deliberately evaluate their tooling based on CI/CD requirements, security constraints, AI integrations, and budget. What are your thoughts? Is your team sticking with GitHub, or have you made the switch to GitLab, Bitbucket, or something entirely different? #SoftwareEngineering #GitHub #GitLab #DevOps #DeveloperProductivity #TechTrends #SoftwareDevelopment #ThePragmaticEngineer

Why I choose Docker Hub for the deployment instead of GitHub 🚀   When it's time to ship code, every developer faces a choice. Do I pull the source code from GitHub? Do I deploy a Docker Image from Docker Hub? For my project, I chose the Docker workflow. Here's why I rely on Docker Desktop and Docker Hub.   The Technical Difference   * GitHub is for storing source code. It has the instructions for your app.   * Docker Hub is for storing container images. It has the environment, like the operating system, libraries, and code, ready to run.   The Breakdown: Docker vs. GitHub for Deployment   Why I Use Docker & Docker Hub:   * I like consistency. By using Docker Desktop to build my image, I know it has every specific library version my app needs.   * This eliminates the problem of "it works on my machine". The image that runs on my laptop is the same one that runs on the server.   However, there is a limitation. Images are larger than code files. So the initial push to Docker Hub takes a bit of bandwidth.   Why I Don't Use GitHub for the Final Deploy:   * GitHub is great for collaboration and version control. It's where my code lives and grows.   *. Deploying straight from GitHub means the server has to build the app. This includes installing dependencies. If a single external library update fails during that build, the whole deployment crashes.   My Deployment Strategy   I use Docker Desktop as my engine. I package everything into a "frozen" container. Pushing this to Docker Hub ensures that:   * Deployment is instant. The server just. Runs. No installing, no compiling.   * Rollbacks are easy. If something goes wrong, I pull an image tag from Docker Hub.   * My infrastructure stays clean. I don't need to install Python, Node, or Java on my server. Just Docker.   I use GitHub to manage my code. I use Docker to manage my runtime. It's the difference between sending someone a recipe and sending them the meal.   Are you Team Docker or Team GitHub?   #Docker #DockerDesktop #DevOps #SoftwareEngineering #DataScience #GitHub #Containerization #TechTalk

💬 Stop Prompting GitHub Copilot. ⚙️ Start Controlling It. The biggest mistake make with GitHub Copilot? Treating it like a tool instead of a system. The real shift now is how you control Copilot using context, commands, and request-based usage. Here’s how GitHub Copilot is actually being used: 🔹 1. Copilot is NOT token-based, it’s request-based This is the most misunderstood part. • 1 prompt you send → counts as 1 request • Copilot generating multiple files → NOT counted • Auto iterations/refinements → NOT counted - Only your prompts are billed If a model has a multiplier (say 3x): 👉 1 prompt = 3 premium requests Free plan: 👉 ~50 premium requests/month Enterprise: 👉 Usage shown in %, controlled by org 💡 This changes how you should use Copilot: → Send fewer, better, well-structured prompts 🔹 2. @ Mentions = Direct Context Injection You don’t need to explain everything anymore. Attach context directly: • @workspace → Works with your entire codebase (best for refactoring, architecture, fixes) • @vscode → Helps with editor features, commands, debugging • @terminal → Understands and works with terminal/CLI output • @github → Brings in PRs, issues, repos, and external knowledge 👉 Copilot works on exact inputs, not assumptions 🔹 3. Slash Commands (/) = Workflow Triggers This is evolving fast. • /explain → understand code • /fix → resolve issues • /tests → generate tests • /doc → create documentation Now expanding into: 👉 Running workflows, fixing pipelines, handling tasks 🔹 4. Context Variables (#) = Precision Control Instead of dumping full code, target exactly: • #selection • #file • #function • #class • #project 👉 You define the scope → Copilot gives better output 🔹 5. Terminal + CLI Integration Copilot is moving closer to where devs actually work: • Execute commands • Fix terminal errors • Stay in flow without switching tabs A common pattern I see is “asking” Copilot for help. It’s in how you interact with it: → Asking questions vs → Sending structured, context-aware requests This is where Copilot starts delivering real value. It’s no longer just assisting, it’s becoming a structured execution layer for development workflows. #GitHubCopilot #AIForDevelopers #DeveloperProductivity #SoftwareEngineering #AI #DevTools #Programming #Coding #TechLeadership #AIEngineering

One of the biggest workflow mistakes I made early in my career was skipping Git and relying on file names, date folders, and server backups. It worked—until it didn’t. Git & GitHub gave me a better way to track changes, recover old work, and collaborate with confidence. For teams, that means fewer mistakes, better accountability, and safer delivery. 💡 #Git #GitHub #VersionControl #DevOps #SoftwareEngineering #TeamCollaboration #TechLeadership https://lnkd.in/dM-MTeQD

devpath-idp update — Phase 6. For those following along: I've been building an internal developer platform from the ground up and sharing the progress here. Phase 5 was software templates. Phase 6 is where things got interesting. One of the trickier debugging lessons I've had so far: A workflow completing successfully doesn't mean it actually did what you think it did. Here's what happened. Phase 6 is about the Backstage scaffolder — the part where a developer fills out a form and it automatically creates a GitHub repo, sets up the structure, and registers the service in the catalog. Self-service provisioning. That's the goal. I ran the template. The scaffolder showed steps completing. No obvious errors in the UI. It looked like it worked. Then I checked GitHub. The repo was empty. Something in the workflow had failed silently. The scaffolder didn't crash — it just didn't finish what it started. The GitHub push step timed out somewhere in the middle, and the UI wasn't loud about it. What made this tricky was the noise around it. When I restarted the backend during earlier sessions, the browser was throwing auth errors and stale token warnings. Those looked serious. They weren't — they were just the browser catching up after a restart. The real failure was quieter and further downstream. That's the thing about debugging distributed workflows: the loudest errors are often not the important ones. The important one here was silent — a repo that existed but had nothing in it. I'm still working through the fix. But the shift in understanding matters: I stopped asking "why is there an error?" and started asking "did this actually complete end to end?" Those are very different questions. And in platform engineering, the second one is usually the right one to ask first. Progress so far: ✅ Phase 1 — Base platform setup ✅ Phase 2 — GitOps foundation ✅ Phase 3 — Backstage portal setup ✅ Phase 4 — Catalog basics ✅ Phase 5 — Software templates 🔧 Phase 6 — Self-service provisioning (in progress) More on this when it's resolved. 🔧 #Backstage #PlatformEngineering #DevOps #InternalDeveloperPlatform #GitHub #Debugging #CloudEngineering #devpath

🚀 Git & GitHub: The Backbone of Modern Software Development Many professionals use Git and GitHub daily, but knowing their history gives better understanding of their importance. 🔹 What is Git ? → Git was created by Linus Torvalds → Year Created: 2005 Git was developed to help the Linux kernel team manage source code changes efficiently. It introduced a fast and distributed version control system that changed the way teams handle code. 🔹 What is GitHub ? → GitHub was founded in: 2008 → Founded by Tom Preston-Werner, Chris Wanstrath, and PJ Hyett GitHub was launched as a cloud platform to host Git repositories and help developers collaborate from anywhere. 🔹 Why Both Became Industry Standards ✅ Git manages version history and code changes ✅ GitHub enables teamwork, remote repositories, code reviews, and project collaboration ✅ Together they became core tools for Developers, DevOps Engineers, and IT teams worldwide 🔹 Major Milestone In 2018, Microsoft acquired GitHub, showing how important GitHub had become in the global tech ecosystem. 🔹 Today’s Reality From startups to global enterprises, Git and GitHub are now part of everyday software development workflows. Some tools are useful. Some tools reshape industries. Git and GitHub did both. #Git #GitHub #DevOps #Linux #AWS #Azure #CloudComputing #Cloud #SoftwareDevelopment #Programming #Developer #Developers #ITJobs #TechJobs #CareerGrowth #OpenToWork #Learning #Technology #Automation #CICD #Jenkins #Docker #Kubernetes #SRE #Infrastructure #SysAdmin #VersionControl #Coding #OpenSource #TechCommunity

🚀 Git vs GitHub - The Most Confusing Duo in Tech Most people think Git is hard… But it really comes down to 5 commands: • git clone → copy a project • git init → start a repo • git add → stage changes • git commit → save changes • git push → upload changes Master these, and you're already ahead of most beginners. 🔹 Git = your local version control system It helps you: • Track every change • Experiment safely with branches • Revert mistakes instantly 🔹 GitHub = Cloud Platform for Collaboration • Review code (Pull Requests) • Collaborate across teams • Run CI/CD pipelines • Track issues and bugs Why this matters: Mastering Git/GitHub isn’t just a “developer/cloud skill". It’s how real engineering teams work: • Work together without breaking things • Ship faster and safer • Build production-grade systems 💡 Git habits that changed everything for me: • Create feature branches (don’t work on main) • Write meaningful commit messages • Use "git stash" when switching tasks • Use "git pull --rebase" for cleaner history • Tag releases (v1.0.0) properly 🔥 If you're learning Git/GitHub right now… You're not behind. You're building the foundation every great engineer stands on. 📘 I’ve attached my Git & GitHub Complete Guide (Beginner → Advanced) Special thanks to Solomon and the entire team at ABC OF CLOUD COMPUTING (CLOUD COMPUTING EMPOWERMENT). Cc: Iberedem, Mmesoma, Bassey-Udofia, Opeyemi, Sarah, Blessing, Ogechukwu, Precious, Situk, Ahmed, Raphael, Nsikan, Uduakabasi, EBENEZER, Joseph, Jude, Jennifer, Divine, Aishat, Felix #Git #GitHub #DevOps #CloudComputing #Azure #SoftwareEngineering #TechCareers #LearnToCode #AWS #Linux #PlatformEngineering #SystemAdministration #LearningInPublic