GitHub Engineer's Accidental rm -rf Disaster: Lessons in System Design and Recovery

GitHub's latest update is a welcome relief for notification fatigue! They're tweaking notification retention periods and, even better, archived repository watches won't be sending you alerts anymore. Finally, some peace for those old, dusty projects! 🙌 A small change, but it really helps keep the focus on what matters. #GitHub #DevOps

Typed “guthib” instead of GitHub today… and landed somewhere I definitely didn’t expect. 😄 It’s funny how in tech, even a single misplaced letter can completely change the outcome—whether it’s a search, a command, or a line of code. Moments like these are small, but they reinforce an important habit: being mindful of the details. Because in our field, precision isn’t optional—it’s everything. Sometimes the best reminders don’t come from big failures, but from tiny slips like this. Back to typing carefully… one character at a time. #DeveloperLife #CodingLessons #AttentionToDetail #TechJourney #GitHub

Github has had 95 incidents in last 90 days. That is more than 1 incident daily. Their SLA promises 99.9% uptime but they have been able to maintain less than 90% That means even if you shut down your product for 1 whole month last year, you still have more uptime than them. Reason is still not clear. But I suspect, since the code commit velocity has increased drastically in last months, The cracks are emerging. Usually with sufficient time between the incidents, dev interventions can fix things. But, at this scale of accelerated usage of product, even the smallest of cracks can become valleys, if not addressed swiftly. Any product's success would depend upon - infrastructure resilience - code quality and most importantly, Proactive foresight while making micro tech decisions. But humans can't foresee everything, we are bound to make mistakes, Which will be visible in our decisions, in our code, and in AI trained on our data. So it's not about avoiding mistakes. Maybe it's about fixing, learning and not repeating What do you think? yeah yeah just let me in🙂 #tech #coding #github #saas

Who knew GitHub settings could trigger a mid-life crisis and a coding lesson at the same time? 😅 The Story: While schecking 2FA on my GitHub today (Because of vercel issue), I stumbled upon the "Successor Settings." GitHub is basically asking: "Who inherits your digital legacy when you're gone?" It hit me differently. As developers and founders, we aren't just writing lines of code; we are building a digital empire. The Realization: If you feel like you aren't "succeeding" today, don't be discouraged. Your work isn't just for this moment. It’s for the next generation. This is a reminder to all my fellow devs: Write Clean Code: Don't leave a mess for your successor. Build logic that is easy to understand. Structure Matters: A well-structured project is the foundation of an empire that stands the test of time. Legacy over Logic: Build things your family and your future generation can be proud of. Let’s stop coding for the weekend and start coding for the legacy. #GitHub #CodingLife #SoftwareEngineering #CleanCode #BuildInPublic #SaaS #Legacy

GitHub is on pace for 14 billion (!) commits this year. Up from one billion in 2025. GitHub's COO, Kyle Daigle, told The Information the platform now sees two hundred and seventy-five million commits a week and is hitting infrastructure limits. Claude Code alone is up twenty-five times in six months. I have been trying to find a historical comparison. Mass manufacturing is the closest I got, because that handed production to machines too. But quality control was easier there and involved statistical sampling on identical units. Code is not identical and every commit is bespoke. The obvious question is whether quality is keeping up? The less obvious one is whether anyone is even checking. My guess is most are not. Fourteen billion commits need fourteen billion answers to the same question. Did this produce any value?

So GitHub now lets orgs disable commit comments. I can already hear the collective sigh of relief from maintainers drowning in long-dead discussions on ancient lines of code. Or maybe a collective groan from those who loved the context? Either way, a powerful new admin lever. What's your take on this cleanup? #GitHub #DevOps

Your PR inbox just got fixed. GitHub rebuilt github.com/pulls from scratch with an Inbox view, saved views, and real filtering. Plus: service container entrypoint overrides in Actions. Also: Simon Willison on why mid-career engineers are the ones actually at risk. Worth the listen. Issue #21: https://lnkd.in/eNJ6Qdvg 🇪🇸 En español · 🇨🇳 中文版 #GitHub #MainBranch #DevTools

THIS. Having GitHub as your only source of truth is a liability waiting to happen. If absolutely ANYTHING happens to your code/account, and you're not paying for GitHub, you're not getting your issues resolved. If you're serious about software and code, owning your remote is way safer than something like GitHub, way more reliable, and not much of a hassle. I've been selfhosting a fork of Forgejo on my home "server", and the memory usage is basically non-existent. Setting up your own Actions runner is also extremely simple, and if you need GitHub Pages, you can either bother messing with Cloudflare Tunnels, or use a public IP address. Setting up a server for it is the less difficult part of it, and TLS is also not that difficult to get. Selfhosting git is a must, and I don't think I can or want to go back. P.S. On the topic of Customer Support, I've also discovered a polar opposite to GitHub's. Brevo's customer support is ACTUAL Customer Support (with, you know, HUMANS reading your tickets?). If you need an SMTP server for transactional emails, I can only recommend.

GitHub celebrated a major milestone in 2025: Achieving 1 billion commits. (A commit represents an update to a codebase.) They're now on pace for 14 billion this year -- and that's if growth stays linear. It won't.

A lot of developers rely on GitHub every single day, but the moment you ask them how it truly differs from GitLab, the answers often get blurry. And honestly, I understand why, on la surface they look similar, yet they don’t serve the same vision at all. GitHub has become the place where the world writes code together. Backed by Microsoft and fueled by a massive open-source community, it’s built for speed, simplicity, and collaboration. Actions, Codespaces, Dependabot… everything is designed to help teams move quickly and stay focused on building. GitLab, on the other hand, follows a completely different philosophy. It’s not just a code platform, it’s a full DevSecOps environment. CI/CD is built-in, security tools are native, governance is centralized, and you can even self-host it with the open-source edition. Many companies choose it because they want one platform to manage everything from planning to deployment. So the question isn’t really “which one is better?”. It’s more like “which vision matches the way you work?”. One focuses on velocity and massive adoption. The other focuses on deep integration and full end-to-end control. If you’ve used either platform in your projects, I’d really love to hear your experience. What actually makes a difference in your daily workflow? And what would you pick again if you had to start from scratch? Your insights will definitely help others who are still trying to choose the right tool. #GitHub #GitLab #DevOps #DevSecOps