Daniel Sogl’s Post

DevOps is more than just "Automation." It’s about Visibility. 🔍 If your team is still treating GitHub as just a "place to put code," you’re missing out on 70% of its value in the DevOps world. True DevOps maturity happens when you leverage: ✅ GitHub Actions for seamless, integrated CI/CD. ✅ GitHub Packages to manage dependencies without leaving the ecosystem. ✅ Environments & Secrets to maintain gatekeeping and security at scale. ✅ Dependabot to handle the "boring" but critical work of version updates. The goal isn't to have more tools; it's to have better integration. By centering the workflow around the repository, we reduce friction and ship faster. GitHub : https://github.com/ #DevOps #PlatformEngineering #GitHubActions #SRE #Automation #Day03

GitHub is starting to feel less like a code repository and more like shared memory for agents. Over the last few days, I cleaned up my workflow around Linear + GitHub + local Obsidian, and the setup is finally starting to make sense: - Obsidian = personal inbox, raw thoughts, fast notes - Linear = tasks, execution, priorities, status - GitHub = the shared knowledge layer for both me and my agents This was the key shift for me: GitHub is now the common storage layer that gives any agent fast access to product context. I’m actively working with 3 different agent providers in parallel, so I need a system where each of them can quickly understand: - strategy - market context - current positioning - custdev learnings - links to execution in Linear - links back to the codebase Instead of re-explaining the same context again and again, I can now point every agent to one structured source of truth. Next step: I want to push this further and set up development more in the direction of Karpathy’s autoresearch idea, where research, synthesis, and execution are connected much more tightly. If you’re working with multiple agents already, I think a shared, structured context layer is becoming mandatory infrastructure, not just a nice-to-have.

🚀 𝗗𝗲𝘃𝗢𝗽𝘀 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗨𝗽𝗱𝗮𝘁𝗲 | 𝗚𝗶𝘁 𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 𝗔𝗰𝗿𝗼𝘀𝘀 𝗠𝘂𝗹𝘁𝗶-𝗥𝗲𝗴𝗶𝗼𝗻 𝗘𝗖𝟮 Continuing my DevOps journey with hands-on practice in Git & GitHub, implementing a real-world workflow across multiple AWS EC2 instances. 🔧 𝗪𝗵𝗮𝘁 𝗜 𝗮𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵𝗲𝗱 𝘁𝗼𝗱𝗮𝘆: ✔ 𝗖𝗿𝗲𝗮𝘁𝗲𝗱 𝗙𝗶𝗿𝘀𝘁 𝗚𝗶𝘁𝗛𝘂𝗯 𝗥𝗲𝗽𝗼𝘀𝗶𝘁𝗼𝗿𝘆 – Set up a central repository to manage code ✔ 𝗟𝗼𝗰𝗮𝗹 𝘁𝗼 𝗥𝗲𝗺𝗼𝘁𝗲 𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 – Initialized local repository and pushed code to GitHub ✔ 𝗠𝘂𝗹𝘁𝗶-𝗥𝗲𝗴𝗶𝗼𝗻 𝗘𝗖𝟮 𝗦𝗲𝘁𝘂𝗽 – Launched 2 EC2 instances in different regions – Simulated distributed environment ✔ 𝗖𝗼𝗱𝗲 𝗦𝘆𝗻𝗰𝗵𝗿𝗼𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻 – Pulled the same code on another EC2 machine – Verified consistency across environments ✔ 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗦𝗲𝘁𝘂𝗽 – Created GitHub Personal Access Token for secure authentication ✔ .𝗴𝗶𝘁𝗶𝗴𝗻𝗼𝗿𝗲 𝗨𝘀𝗮𝗴𝗲 – Excluded unnecessary files from version control 💻 𝗖𝗼𝗺𝗺𝗮𝗻𝗱𝘀 𝗨𝘀𝗲𝗱 & 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴: git init ➡ Initializes a new local Git repository git add . ➡ Stages all files for commit git status ➡ Shows current state of working directory git commit -m "First commit from Mumbai" ➡ Saves changes with a message git log ➡ Displays commit history git show ➡ Shows details of a specific commit git remote add origin https://lnkd.in/d9x5Rwpz ➡ Connects local repo to GitHub repository git push origin master ➡ Pushes code to GitHub (central repo) git pull origin master ➡ Fetches and updates code from GitHub 💡 𝗞𝗲𝘆 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆: Understanding Git commands along with real environment practice (multi-EC2 setup) builds a strong foundation for collaboration and deployment in DevOps. Learning by doing—step by step towards becoming a DevOps Engineer 🚀 #DevOps #Git #GitHub #AWS #EC2 #VersionControl #CloudComputing #HandsOnLearning #LearningJourney

GitHub is bringing stacked PRs natively into the platform — the ability to break large changes into a clean chain of small, focused, reviewable pull requests that build on each other. Stacked diffs aren’t new — they’ve been battle-tested for years at Google, Meta, and other FAANG companies. The real value is velocity: breaking big changes into small, incremental pieces that ship fast instead of getting stuck in one massive PR. This approach finally helps teams shake off those old SVN-era habits — giant monolithic diffs that were painful to review, full of conflicts, and slowed everything down. With stacked PRs, you keep momentum, get faster feedback, and land changes efficiently. You can create a chain where each PR targets the branch of the one below it. Everything merges smoothly into main, and reviewers get proper context without drowning in huge diffs. Graphite built some genuinely impressive stacked diff tooling, but I suspect most teams will lean toward GitHub’s built-in experience as it matures. We’ll see what creative open-source solutions the community builds to tidy up the workflow. What stands out: 🡆 A clean stack map in the GitHub PR UI for easy navigation between layers 🡆 Automatic cascading rebases — merge one and the rest of the stack updates intelligently 🡆 Full branch protection and CI enforcement on the final target branch 🡆 The new gh stack CLI that makes creating layers, pushing the entire stack, submitting all PRs at once, and handling rebases seamless 🡆A clean linear change history! No slop commits! That said, having to rely on an external CLI tool — unlike Gerrit’s more integrated approach — still feels a bit unfortunate. This feels like the end-to-end developer workflow finally coming together inside GitHub itself. For most teams already living in GitHub, the convenience and tight integration will be very hard to beat. This goes some way toward GitHub repairing the anti-patterns they have perpetuated — hopefully their next trick will be doing something to curb repo sprawl. ↳ https://lnkd.in/gYNeaK8T

🔄 Day 9 of the 30-Day Terraform Challenge — went deep on advanced Terraform modules today. Module versioning, file path gotchas, and deploying different module versions across dev and production. This is the pattern that keeps large infrastructure codebases manageable. Here is the link to the Blog: https://lnkd.in/dmBaj9aB #30DayTerraformChallenge #TerraformChallenge #Terraform #IaC #DevOps #AWSUserGroupKenya #EveOps

Pushing code is easy… Making sure it works every time is hard 😅 That’s where GitHub Actions comes in. It lets you automate your workflow: • Build your code • Run tests • Deploy automatically All triggered by events like push or PR 🚀 Read more here: https://lnkd.in/gmUXX5MM - By Octopus Deploy

PR is ready. Tests are passing. …and then it just sits there👀 Because no one wants to pick who reviews it 😅 If you’ve worked in a team, you’ve probably seen this. So I built a small GitHub Action to fix it: 👉 https://lnkd.in/dD_BTWhF What it does: • Pulls team members from your GitHub org • Excludes the PR author • Randomly picks one reviewer • Automatically requests the review via GitHub API Why I built it: • Remove the manual “who’s reviewing this?” step • Distribute reviews more fairly across the team • Keep PRs moving (without Slack nudges 👀) It’s intentionally simple, but plugs directly into existing workflows (pull_request trigger, supports GITHUB_TOKEN / PAT). Sometimes the best improvements are the small ones that remove daily friction #SoftwareEngineering #GitHub #CodeReview #Automation #DeveloperExperience #OpenSource

I just stumbled across gh-stack, and I’m curious. is anyone actually using this in their daily workflow? 🧐 It’s a GitHub CLI extension designed for "Stacked Pull Requests." The idea is to break down massive changes into a series of small, dependent PRs that can be reviewed independently, but managed as a single "stack." The highlights that caught my eye: ✅ It automates the messy rebasing that usually happens when you stack PRs. ✅ It handles the PR "base branch" logic automatically on GitHub. ✅ It keeps the PRs small and reviewable (the dream!). I’ve always found managing stacks manually in Git to be a bit of a nightmare. This looks like it might actually make the process painless. Has anyone tried this? Does it live up to the hype, or are there better alternatives out there? 🔗 https://lnkd.in/e4NMZHAB #GitHub #SoftwareDevelopment #Git #DeveloperTools #DeveloperExperience

Just shipped my latest open source project: deploycheck 🚀 It's a Claude Code plugin that lets you ask "is staging ready to deploy?" - and actually get a real answer. 🎉 Instead of switching between GitHub Actions, Docker Hub, and your health dashboards before every deploy, Claude runs the checks for you: → Health endpoint checks across all environments 🐳 → CI pipeline status (GitHub Actions / GitLab CI) ✅ → Full readiness sweep with a go/no-go recommendation 🚦 🔗 repo: https://lnkd.in/dvwjawgQ If you're working with MCP or building Claude plugins, I'd love to hear what you think. #claude #claudecode #opensource #claudeplugin #mcp

The GitHub algorithm is just another system to reverse-engineer. 🟩⚡ **Today’s drop: The Commits Cheat Code.** I showed exactly how to manipulate Git history and paint your contribution graph solid green. Is it a vanity metric? Absolutely. But understanding how Git actually handles timestamps, staging, and commit hashes isn't just a party trick—it’s a deep dive into how version control architecture actually works under the hood. A massive green wall looks premium on a profile, but here is the reality check: the squares might get you the initial attention, but the actual logic inside the repositories is what closes the client or gets you the job. Use the trick to fix a broken streak. But don't forget to build real systems. 🏗️💻 Are you obsessed with keeping your daily Git streak alive, or do you only push when the feature is completely finished? 👇 #GitHub #Git #SoftwareEngineering #WebDevelopment #CodingHacks #100DaysOfCode #DeveloperLife #TechAesthetic #BuildInPublic #TechMinimalist