Hell yeah! Copilot on self-hosted runners. Fast and cost efficient infrastructure for runners has never been more important than now. #GitHubUniverse

Who is in the city for #GitHubUniverse? 🌁 Let's talk DevTools and specifically about ephemeral environments for your CI/CD workloads and of course AI agents to work in.

📰 Press Release: we successfully enforced our Fair Source License for Tart Virtualization. This was an exceptional case since we explicitly declined a license to this company due to a conflict of interest. Read more in the post below 👇 https://t.co/gX7dh5x6z9

See you all in Amsterdam next year! 🌷🇳🇱

Have they started moving @GitHub over to @Azure already? @githubstatus is celebrating with a firework of colors! 🎉🚒

Great presentation by engineers @Jamf on how they use Tart and how composable Tart and all the tools around are. We always aimed for such composability and happy to see usage outside of CI/CD space. PS in love with the baking theme of the talk! 🧑‍🍳👩‍🍳 https://t.co/D3fyAKaoQH

Good morning, @NSSpain! Come to our booth to talk about virtualization and running CI/CD as fast as possible. 🤓

Good morning! This week marks new Liquid Glass UI releases over iOS and macOS! 🎉 To celebrate the release we are introducing support for GitHub Actions Runners based of the macOS Tahoe RC with the latest Xcode. Try it with a one line change 👇 runs-on: https://t.co/HNXg7XfR9g

CI configs but Starlark instead of YAML. 🚀 Oh... wait. We did it!

Good time to try alternatives? Our Linux runners are twice as fast 20 times cheaper in some case.

This is why our big enterprise customers like @Zendesk see not 2x but 5 to 10 times smaller bills for GitHub Actions Runners. Thank you for reading till the end! We don't have a permanent free trier hence no VC money but our free trial is two weeks. Please give it a shot!

Image a tradition per-minute provider backed by a VC will make all builds 40% faster without increasing prices. It means immediate revenue drop by 40%. This is along the lines of conversations by @levelsio and @dhh about hosting infrastructure off traditional clouds.

When we make our infrastructure faster we don't charge more. It's been always a fixed $150 per runner for unlimited minutes. This is not fueled by VC money! We are bootstrapped.

This all paid off. We are now #1 provider for GitHub Actions in terms of raw performance according to an independent benchmarks. See blog post for details 👇 https://t.co/eoSdUJ0yqB

Then Linux x86 VMs are executed on high-frequency @AMDRyzen CPUs (think of gaming CPUs but with server reliability). Linux Arm VMs are actually still run on M4 Pro Apple Silicon chips. They are almost 3 times faster then server @Arm chips like @AmpereComputing.

We developed two virtualizations to run the same images on Apple Silicon chips and traditional Linux servers. We build all Linux images on Apple Silicon and Intel Mac Minis and publish them to a Docker registry in a generic format.

🧵Did you know that Cirrus Runners is suitable not only for mobile CI/CD? We started with a focus on iOS devs first and then expanded to Android teams. Now we conquer tradition Linux workloads too! Around 75% of the GHA jobs we run is Linux nowadays. See distribution below 👇

Merged PR from willcl-ark: ci Migrate CI to hosted Cirrus Runners

Or inside of a macOS VM via https://t.co/58JEWFyJdM 🤓🍏

Do you run GitHub self-hosted runners? How do you manage them in production across provisioning, autoscaling, job isolation, secrets, and upgrades? How many runners do you run in parallel, and what is your monthly all-in cost for compute, storage, and ops?