# Avrea Blog

All the latest and greatest from the land of CI/CD.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Avrea Sandboxes: computers should be forkable

DevFeed: [Avrea Sandboxes: computers should be forkable](<https://devfeed.tech/articles/avrea-sandboxes-computers-should-be-forkable-132348.md>)

Original publisher: [Read original article](<https://avrea.com/blog/avrea-sandboxes-computers-should-be-forkable>)

Author: Markus Stenberg

Published: 2026-11-13T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [virtual machines](<https://devfeed.tech/topics/virtual-machines.md>), [virtualization](<https://devfeed.tech/topics/virtualization.md>), [self-service](<https://devfeed.tech/topics/self-service.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [x86](<https://devfeed.tech/topics/x86.md>), [pnpm](<https://devfeed.tech/topics/pnpm.md>)

Tags: [1-4](<https://devfeed.tech/tags/1-4.md>), [2](<https://devfeed.tech/tags/2.md>), [about](<https://devfeed.tech/tags/about.md>), [abstraction](<https://devfeed.tech/tags/abstraction.md>), [avrea](<https://devfeed.tech/tags/avrea.md>), [billing](<https://devfeed.tech/tags/billing.md>), [can](<https://devfeed.tech/tags/can.md>), [computers](<https://devfeed.tech/tags/computers.md>), [fork](<https://devfeed.tech/tags/fork.md>), [full](<https://devfeed.tech/tags/full.md>), [git](<https://devfeed.tech/tags/git.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [https](<https://devfeed.tech/tags/https.md>), [m5-max](<https://devfeed.tech/tags/m5-max.md>), [max](<https://devfeed.tech/tags/max.md>), [mirrors](<https://devfeed.tech/tags/mirrors.md>), [package](<https://devfeed.tech/tags/package.md>), [sandboxes](<https://devfeed.tech/tags/sandboxes.md>), [with](<https://devfeed.tech/tags/with.md>), [zen-5](<https://devfeed.tech/tags/zen-5.md>)

### AI overview

Avrea Sandboxes are full virtual machines intended as temporary computers for developers and agents, with hardware virtualization, non-oversubscribed CPUs, and Linux arm64 M5 Max or Linux amd64 and Windows Zen 5 hardware. Datacenter-local Git mirrors and colocated package caches reduce setup work; repository-connected sandboxes can also inherit configured CI caches. They support per-second billing, included networking, and public HTTPS endpoints with Avrea-managed TLS. Eligible paused VMs can be forked into independent copies of disk and machine configuration, but not external side effects or running process memory. Forking and pause/resume currently require supported paused sandboxes, with pause/resume support limited to Linux amd64; the separate ephemeral agent-sandbox service does not support forking. Published URLs are public and sensitive applications still need authentication. Performance and build figures cited are workload-specific.

### Source excerpt

Avrea Sandboxes are full VMs on M5 Max and Zen 5 hardware, with Git mirrors, package caches, HTTPS endpoints and per-second billing. And you can fork them.

## StepSecurity Harden-Runner now supports Avrea runners

DevFeed: [StepSecurity Harden-Runner now supports Avrea runners](<https://devfeed.tech/articles/stepsecurity-harden-runner-now-supports-avrea-runners-122865.md>)

Original publisher: [Read original article](<https://avrea.com/blog/stepsecurity-harden-runner>)

Author: Tatu Mäkijärvi

Published: 2026-11-06T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Endpoint security](<https://devfeed.tech/topics/endpoint-security.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Process](<https://devfeed.tech/topics/process.md>), [macOS](<https://devfeed.tech/topics/macos.md>)

Tags: [linux](<https://devfeed.tech/tags/linux.md>), [macos](<https://devfeed.tech/tags/macos.md>), [network](<https://devfeed.tech/tags/network.md>), [process](<https://devfeed.tech/tags/process.md>), [records](<https://devfeed.tech/tags/records.md>), [runner](<https://devfeed.tech/tags/runner.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

Harden-Runner v2.22.3 or later runs on Avrea's x86_64 and ARM64 Linux runners and macOS. Added as the first step of each job, its agent records network connections and process activity for inspection by workflow step in StepSecurity's dashboard. Avrea's egress firewall shows job connections but not which step or process initiated them; Harden-Runner adds that detail. With `egress-policy: audit`, it records outbound traffic without enforcing an allowlist, while StepSecurity's Global Block List still blocks known malicious destinations. Avrea-specific settings are not required. Access requires a StepSecurity Enterprise plan or trial; the free tier covers public repositories on GitHub-hosted runners, and Avrea counts as self-hosted. Verify monitoring by checking for network and process events in the dashboard: the action can pass even if no agent was installed. Avrea runner service connections appear separately as Infrastructure Events. Windows runners are unsupported.

### Source excerpt

StepSecurity's Harden-Runner now runs on Avrea's Linux and macOS runners and records each workflow step's network connections and process activity.

## Avrea product update, October 2026

DevFeed: [Avrea product update, October 2026](<https://devfeed.tech/articles/avrea-product-update-october-2026-122839.md>)

Original publisher: [Read original article](<https://avrea.com/blog/avrea-product-update-october-2026>)

Author: Tatu Mäkijärvi

Published: 2026-11-04T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Release notes](<https://devfeed.tech/topics/release-notes.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [actions](<https://devfeed.tech/tags/actions.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [arm](<https://devfeed.tech/tags/arm.md>), [assets](<https://devfeed.tech/tags/assets.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cache](<https://devfeed.tech/tags/cache.md>), [changelog](<https://devfeed.tech/tags/changelog.md>), [downloads](<https://devfeed.tech/tags/downloads.md>), [graph](<https://devfeed.tech/tags/graph.md>), [half](<https://devfeed.tech/tags/half.md>), [linux](<https://devfeed.tech/tags/linux.md>), [memory](<https://devfeed.tech/tags/memory.md>), [october](<https://devfeed.tech/tags/october.md>), [october-2026](<https://devfeed.tech/tags/october-2026.md>), [product](<https://devfeed.tech/tags/product.md>), [product-update](<https://devfeed.tech/tags/product-update.md>), [right-sizing](<https://devfeed.tech/tags/right-sizing.md>), [run](<https://devfeed.tech/tags/run.md>), [runner](<https://devfeed.tech/tags/runner.md>)

### AI overview

Avrea's October 2026 update adds runner sizing guidance, persistent disks, workflow visualization, and more cached downloads. Right-sizing compares each job's recent successful runs with its CPU, memory, and disk use, then suggests a runner label; a workflow-wide view and copyable coding-assistant prompt are available to every organization. Persistent disks keep repository files between jobs on Linux, macOS, and Windows: default-branch jobs publish changes, while pull-request and other-branch changes are discarded. Disks are paid per provisioned GB and must be enabled in repository settings. Two lower-rate Linux sizes retain 4 or 8 vCPUs with half the memory and are available on x86 and ARM. The run graph shows job dependencies, status, duration, steps, logs, and a timeline of queued, running, and completed work. Cached Homebrew, mise, and GitHub release downloads are off by default; Homebrew and mise require Package Manager Cache. Electron, Cypress, and Selenium downloads are enabled by default. Checksum and signature verification remain enabled.

### Source excerpt

Runner right-sizing, persistent disks, Linux runners with half the memory, a run graph and more cached downloads: what shipped in Avrea in October 2026.

## Avrea is the fastest GitHub Actions runner in RunsOn's benchmark

DevFeed: [Avrea is the fastest GitHub Actions runner in RunsOn's benchmark](<https://devfeed.tech/articles/avrea-is-the-fastest-github-actions-runner-in-runson-s-benchmark-122840.md>)

Original publisher: [Read original article](<https://avrea.com/blog/avrea-runson-build-benchmark>)

Author: Hannu Valtonen

Published: 2026-11-03T00:00:00Z

Content type: news

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [Bun](<https://devfeed.tech/topics/bun.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [app](<https://devfeed.tech/tags/app.md>), [backend](<https://devfeed.tech/tags/backend.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [build](<https://devfeed.tech/tags/build.md>), [build-times](<https://devfeed.tech/tags/build-times.md>), [builds](<https://devfeed.tech/tags/builds.md>), [buildx](<https://devfeed.tech/tags/buildx.md>), [bun](<https://devfeed.tech/tags/bun.md>), [c](<https://devfeed.tech/tags/c.md>), [c-plus-plus](<https://devfeed.tech/tags/c-plus-plus.md>), [cache](<https://devfeed.tech/tags/cache.md>), [docker](<https://devfeed.tech/tags/docker.md>), [faster](<https://devfeed.tech/tags/faster.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [results](<https://devfeed.tech/tags/results.md>), [runner](<https://devfeed.tech/tags/runner.md>), [rust](<https://devfeed.tech/tags/rust.md>), [typescript](<https://devfeed.tech/tags/typescript.md>)

### AI overview

RunsOn's benchmark reports that Avrea's four-vCPU x64 runner completed three uncached builds faster and at lower estimated cost than GitHub's: PostHog's Docker build took 4m52s versus 12m53s, opencode's TypeScript checks 40s versus 1m52s, and spotify-player's Rust job 1m26s versus 3m33s. RunsOn sums median build times for its ranking and publishes methodology, code, and logs. For these three builds, it estimates $0.0625 on Avrea versus $0.2612 on GitHub; the per-minute rates cited are $0.008 and $0.012, respectively. The excerpt says Avrea supports Turborepo remote caching, sccache for Rust/C/C++, and colocated dependency caches for npm, pnpm, Yarn Classic, and Bun. It suggests trying the same workflow and commit on both providers, then rerunning on Avrea to observe caching effects. Results reflect the specified workloads and configurations.

### Source excerpt

Avrea leads RunsOn's 4-vCPU x64 build benchmark, completing Rust, TypeScript and Docker workloads 163% faster than GitHub. See all provider results.

## Warm builds. Introducing Avrea persistent disks

DevFeed: [Warm builds. Introducing Avrea persistent disks](<https://devfeed.tech/articles/warm-builds-introducing-avrea-persistent-disks-122859.md>)

Original publisher: [Read original article](<https://avrea.com/blog/persistent-disks>)

Author: Hannu Valtonen

Published: 2026-10-27T10:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Filesystems](<https://devfeed.tech/topics/filesystems.md>), [build performance](<https://devfeed.tech/topics/build-performance.md>), [container images](<https://devfeed.tech/topics/container-images.md>), [builds](<https://devfeed.tech/topics/builds.md>)

Tags: [also](<https://devfeed.tech/tags/also.md>), [application](<https://devfeed.tech/tags/application.md>), [archive](<https://devfeed.tech/tags/archive.md>), [automatically](<https://devfeed.tech/tags/automatically.md>), [avoid](<https://devfeed.tech/tags/avoid.md>), [away](<https://devfeed.tech/tags/away.md>), [based](<https://devfeed.tech/tags/based.md>), [benefit](<https://devfeed.tech/tags/benefit.md>), [best](<https://devfeed.tech/tags/best.md>), [branch](<https://devfeed.tech/tags/branch.md>), [build](<https://devfeed.tech/tags/build.md>), [builds](<https://devfeed.tech/tags/builds.md>), [ci](<https://devfeed.tech/tags/ci.md>), [docker](<https://devfeed.tech/tags/docker.md>), [jobs](<https://devfeed.tech/tags/jobs.md>), [keep](<https://devfeed.tech/tags/keep.md>), [layers](<https://devfeed.tech/tags/layers.md>), [reuse](<https://devfeed.tech/tags/reuse.md>), [with](<https://devfeed.tech/tags/with.md>), [work](<https://devfeed.tech/tags/work.md>)

### AI overview

Avrea persistent disks keep selected cache files across fresh CI VMs, letting tools reuse files in place instead of restoring and uploading archives. Teams choose cache paths and writer branches; jobs use separate copies, and pull requests can read saved state while their changes are discarded. Concurrent writers do not merge: the first to finish publishes, while stale writers are discarded. Normally completed writer jobs publish even after a step fails; cancelled or crashed jobs do not. Linux Docker caching can preserve images, build cache and volumes; container-based Buildx builders require a stable name and `keep-state: true`. Disks are best effort, may be unavailable in some datacenters, and jobs must work with an empty cache. Provisioned capacity is billed by OS and datacenter at $0.50/GB/month, whether full or not; the Docker cache uses the Linux disk at no extra charge.

### Source excerpt

Keep build caches and Docker layers between CI jobs with Avrea persistent disks. Reuse the work already done between your jobs.

## Avrea is the fastest ARM runner for GitHub Actions: 2026 benchmark

DevFeed: [Avrea is the fastest ARM runner for GitHub Actions: 2026 benchmark](<https://devfeed.tech/articles/avrea-is-the-fastest-arm-runner-for-github-actions-2026-benchmark-122843.md>)

Original publisher: [Read original article](<https://avrea.com/blog/fastest-arm-github-actions-runners>)

Author: Tatu Mäkijärvi

Published: 2026-10-23T00:00:00Z

Content type: news

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [GitHub](<https://devfeed.tech/topics/github.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-m5](<https://devfeed.tech/tags/apple-m5.md>), [arm](<https://devfeed.tech/tags/arm.md>), [arm64](<https://devfeed.tech/tags/arm64.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [build](<https://devfeed.tech/tags/build.md>), [cargo](<https://devfeed.tech/tags/cargo.md>), [ci](<https://devfeed.tech/tags/ci.md>), [fast](<https://devfeed.tech/tags/fast.md>), [faster](<https://devfeed.tech/tags/faster.md>), [free-tier](<https://devfeed.tech/tags/free-tier.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [m5-max](<https://devfeed.tech/tags/m5-max.md>), [runner](<https://devfeed.tech/tags/runner.md>)

### AI overview

In RunsOn's November 2026 cold Rust build benchmark, Avrea's Apple M5 Max ARM runners finished first among 36 runners from seven providers at 2, 4 and 8 vCPU, taking 116, 76 and 58 seconds. GitHub's ARM runners took 316 and 196 seconds at 2 and 4 vCPU; Avrea was 2.7x and 2.6x as fast at those sizes. The benchmark used median build-step times from five cold runs of spotify-player with no cache. RunsOn reports higher arm64 single-thread PassMark scores for Avrea's M5 Max, a factor it links to Rust's serial compilation path. Estimated full-job cost at 2 vCPU was $0.025 on Avrea versus $0.030 on GitHub after GitHub's whole-minute rounding, before allowances; GitHub-hosted runners are free for public repositories. To use Avrea ARM runners, install its GitHub App and choose a runner label. Ubuntu 24.04 and 26.04 sizes range from 1 to 16 vCPU; the free tier covers about 500 minutes on the 4-vCPU, 16-GB runner, shared with other usage.

### Source excerpt

RunsOn's public benchmark of 36 arm64 GitHub Actions runners from seven providers: Avrea's M5 Max runners finished first at 2, 4 and 8 vCPU.

## Runner right-sizing: stop paying for capacity your jobs don't use

DevFeed: [Runner right-sizing: stop paying for capacity your jobs don't use](<https://devfeed.tech/articles/runner-right-sizing-stop-paying-for-capacity-your-jobs-don-t-use-122861.md>)

Original publisher: [Read original article](<https://avrea.com/blog/runner-right-sizing>)

Author: Tatu Mäkijärvi

Published: 2026-10-22T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [internal developer platform automation](<https://devfeed.tech/topics/internal-developer-platform-automation.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [costs](<https://devfeed.tech/tags/costs.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [memory](<https://devfeed.tech/tags/memory.md>), [right-sizing](<https://devfeed.tech/tags/right-sizing.md>), [runner](<https://devfeed.tech/tags/runner.md>)

### AI overview

Avrea recommends runner sizes per job from CPU, memory and disk use across recent successful runs, naming a label to switch to. After three successful runs on a size, jobs consistently below capacity may get a Size down suggestion; high average CPU or memory or disk peaks near capacity may prompt Size up. Suggestions appear on job and run pages, in a repository workflow view covering the last day, 7 days or 30 days, and through the API. A full persistent disk is flagged separately because changing runners does not expand it. Since CPU is assessed as a run average, brief intensive work after long idle periods may slow down on a smaller runner; the new size is evaluated through subsequent runs.

### Source excerpt

Avrea now suggests a smaller or larger runner for each job, based on the CPU, memory and disk its recent runs used, and gives you the label to switch to.

## Get to the build sooner: Git pre-checkout, LFS caching, and submodules

DevFeed: [Get to the build sooner: Git pre-checkout, LFS caching, and submodules](<https://devfeed.tech/articles/get-to-the-build-sooner-git-pre-checkout-lfs-caching-and-submodules-122846.md>)

Original publisher: [Read original article](<https://avrea.com/blog/git-checkout-acceleration>)

Author: Hannu Valtonen

Published: 2026-10-16T09:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Git](<https://devfeed.tech/topics/git.md>), [Submodules](<https://devfeed.tech/topics/submodules.md>), [Dependency management](<https://devfeed.tech/topics/dependency-management.md>), [build times](<https://devfeed.tech/topics/build-times.md>)

Tags: [acceleration](<https://devfeed.tech/tags/acceleration.md>), [actions](<https://devfeed.tech/tags/actions.md>), [all](<https://devfeed.tech/tags/all.md>), [app](<https://devfeed.tech/tags/app.md>), [asset](<https://devfeed.tech/tags/asset.md>), [assets](<https://devfeed.tech/tags/assets.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [avr](<https://devfeed.tech/tags/avr.md>), [benefit](<https://devfeed.tech/tags/benefit.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [build](<https://devfeed.tech/tags/build.md>), [caching](<https://devfeed.tech/tags/caching.md>), [can](<https://devfeed.tech/tags/can.md>), [checkout](<https://devfeed.tech/tags/checkout.md>), [ci](<https://devfeed.tech/tags/ci.md>), [downloads](<https://devfeed.tech/tags/downloads.md>), [git](<https://devfeed.tech/tags/git.md>), [large](<https://devfeed.tech/tags/large.md>), [repositories](<https://devfeed.tech/tags/repositories.md>), [submodules](<https://devfeed.tech/tags/submodules.md>)

### AI overview

Avrea speeds CI startup by preparing Git objects before checkout, caching Git LFS downloads near its runners, and preloading pinned submodule commits. `actions/checkout` remains in place and reuses prepared objects; unavailable mirrors and cache misses fall back to GitHub. Git Checkout Acceleration and Git LFS Cache are off by default, while submodule preparation defaults to root-only. Teams can choose top-level or recursive submodules; private submodules require GitHub App authorization, a ready mirror, and the credentials ordinary checkout needs. Root and submodule pre-checkout support GitHub repositories on Linux and macOS; Windows uses normal checkout, and LFS downloads inside containers are not intercepted.

### Source excerpt

Avrea prepares Git repositories before checkout, caches large LFS downloads, and preloads submodules so CI can get to the build sooner.

## GitHub Actions pricing in 2026 and how to lower your bill

DevFeed: [GitHub Actions pricing in 2026 and how to lower your bill](<https://devfeed.tech/articles/github-actions-pricing-in-2026-and-how-to-lower-your-bill-122849.md>)

Original publisher: [Read original article](<https://avrea.com/blog/github-actions-pricing>)

Author: Tatu Mäkijärvi

Published: 2026-10-02T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [GitHub](<https://devfeed.tech/topics/github.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [actions](<https://devfeed.tech/tags/actions.md>), [bill](<https://devfeed.tech/tags/bill.md>), [billing](<https://devfeed.tech/tags/billing.md>), [budget](<https://devfeed.tech/tags/budget.md>), [build](<https://devfeed.tech/tags/build.md>), [caching](<https://devfeed.tech/tags/caching.md>), [compute](<https://devfeed.tech/tags/compute.md>), [costs](<https://devfeed.tech/tags/costs.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [pricing](<https://devfeed.tech/tags/pricing.md>), [runner](<https://devfeed.tech/tags/runner.md>)

### AI overview

GitHub Actions charges private-repository usage beyond plan compute allowances by runner rate and per-job runtime rounded up to a whole minute; failed and rerun jobs also count. Standard public-repository runners and self-hosted runner usage have no Actions fee, while storage has separate rules. The text identifies practical savings: inspect billing by SKU and repository, combine very short jobs, put cheap checks before expensive ones, cancel obsolete runs, set timeouts, and choose the least costly suitable runner. Parallel shards can shorten waits but repeated setup and per-job rounding can raise charges. It compares third-party providers, especially Avrea, whose described pricing bills seconds rounded down with no minimum and whose benchmarks report faster builds and cache restores; results vary by workload, and jobs waiting on external services may not speed up. The text also notes Avrea ARM rates can be higher than GitHub's, requiring enough speedup to offset the difference.

### Source excerpt

What GitHub Actions costs in 2026: per-minute rates for every hosted runner, how the bill is calculated, and the settings and runner choices that lower it.

## iOS and macOS CI benchmark: Avrea is 3-23x faster than GitHub Actions

DevFeed: [iOS and macOS CI benchmark: Avrea is 3-23x faster than GitHub Actions](<https://devfeed.tech/articles/ios-and-macos-ci-benchmark-avrea-is-3-23x-faster-than-github-actions-122853.md>)

Original publisher: [Read original article](<https://avrea.com/blog/ios-macos-ci-benchmark>)

Author: Leo Lännenmäki

Published: 2026-10-01T00:00:00Z

Content type: comparison

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [macOS](<https://devfeed.tech/topics/macos.md>), [iOS](<https://devfeed.tech/topics/ios.md>), [mobile devops](<https://devfeed.tech/topics/mobile-devops.md>)

Tags: [apple](<https://devfeed.tech/tags/apple.md>), [apple-m5](<https://devfeed.tech/tags/apple-m5.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [build](<https://devfeed.tech/tags/build.md>), [cache](<https://devfeed.tech/tags/cache.md>), [caching](<https://devfeed.tech/tags/caching.md>), [ci](<https://devfeed.tech/tags/ci.md>), [clang](<https://devfeed.tech/tags/clang.md>), [commit](<https://devfeed.tech/tags/commit.md>), [commits](<https://devfeed.tech/tags/commits.md>), [compilation](<https://devfeed.tech/tags/compilation.md>), [dependencies](<https://devfeed.tech/tags/dependencies.md>), [dependency](<https://devfeed.tech/tags/dependency.md>), [firefox](<https://devfeed.tech/tags/firefox.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [ios](<https://devfeed.tech/tags/ios.md>), [macos](<https://devfeed.tech/tags/macos.md>), [mastodon](<https://devfeed.tech/tags/mastodon.md>), [tuist](<https://devfeed.tech/tags/tuist.md>)

### AI overview

Avrea's Apple M5 Max runners and colocated caches cut build times and per-build costs in the tested projects. Firefox for iOS took just under 2 minutes with a warm cache versus over 12 minutes on GitHub's standard macOS runner; Mastodon built with Tuist took 11.8 seconds versus 269.8 seconds. Without caches, the M5 Max was 3.6-6.9 times faster than GitHub's standard runner. Avrea combines Xcode compilation caching, Tuist's prebuilt-target module cache, and a Swift package registry; the registry speeds dependency resolution rather than compilation. Results vary: Firefox's generated timestamps invalidate cache keys unless pinned, Signal disables caching for its explicitly built modules, and small or non-compilation-heavy builds gain less. Across the tested build steps, Avrea cost 3.5-20 times less per build than GitHub's standard runner, despite costing 29% more per minute. Measurements used pinned commits and warm caches for Avrea, while GitHub results had no cache; registry speedups were reported only for Vapor and The Composable Architecture because resolution failed for three tested apps.

### Source excerpt

On GitHub Actions' standard macOS runner, a Firefox for iOS build takes over 12 minutes and on Avrea with a warm cache it takes just under 2 minutes. Mastodon, built with Tuist, goes from 4 minutes 30 seconds to 12 seconds.

## GitHub Actions observability: understanding why your CI is slow

DevFeed: [GitHub Actions observability: understanding why your CI is slow](<https://devfeed.tech/articles/github-actions-observability-understanding-why-your-ci-is-slow-122848.md>)

Original publisher: [Read original article](<https://avrea.com/blog/github-actions-observability>)

Author: Tatu Mäkijärvi

Published: 2026-09-16T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [observability](<https://devfeed.tech/topics/observability.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [OpenTelemetry](<https://devfeed.tech/topics/opentelemetry.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [actions](<https://devfeed.tech/tags/actions.md>), [added](<https://devfeed.tech/tags/added.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [all](<https://devfeed.tech/tags/all.md>), [cache](<https://devfeed.tech/tags/cache.md>), [ci](<https://devfeed.tech/tags/ci.md>), [data](<https://devfeed.tech/tags/data.md>), [flaky](<https://devfeed.tech/tags/flaky.md>), [flaky-tests](<https://devfeed.tech/tags/flaky-tests.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [jobs](<https://devfeed.tech/tags/jobs.md>), [logs](<https://devfeed.tech/tags/logs.md>), [metrics](<https://devfeed.tech/tags/metrics.md>), [observability](<https://devfeed.tech/tags/observability.md>), [runner](<https://devfeed.tech/tags/runner.md>), [tests](<https://devfeed.tech/tags/tests.md>), [timing](<https://devfeed.tech/tags/timing.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

Investigate CI delays by separating queue time from job execution, comparing equivalent runs, and lining step timings up with logs and runner measurements. GitHub provides workflow and step durations, logs, runner image details, aggregate queue and usage metrics, job timestamps through its API, and targeted debug logging; access to aggregate views depends on repository or organization permissions. Retained CPU, memory, disk, and network data can help distinguish resource pressure from slow commands. Cache hits do not guarantee faster builds: package caching may avoid downloads while installs or compilation still run, and cache transfer can cost more than regeneration. Preserve individual test attempts and conditions because a green rerun or internally retried test can hide an initial failure. Avrea adds per-job resource metrics, cache entry details, run history and trends, failed-step log retrieval, live job monitoring, and SSH inspection before the VM is removed; its flake metric is a lead to investigate, not proof of a flaky test. OpenTelemetry is optional and most useful when CI data must join an existing telemetry platform or span multiple CI systems.

### Source excerpt

Use GitHub Actions logs, workflow timing, runner metrics, and cache data to investigate slow jobs, flaky tests, and CI delays.

## A guide to GitHub Actions caching

DevFeed: [A guide to GitHub Actions caching](<https://devfeed.tech/articles/a-guide-to-github-actions-caching-122847.md>)

Original publisher: [Read original article](<https://avrea.com/blog/github-actions-cache>)

Author: Tatu Mäkijärvi

Published: 2026-09-01T00:00:00Z

Content type: tutorial

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Cache](<https://devfeed.tech/topics/cache.md>), [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [GitHub](<https://devfeed.tech/topics/github.md>), [GitHub API](<https://devfeed.tech/topics/github-api.md>), [pnpm](<https://devfeed.tech/topics/pnpm.md>)

Tags: [backends](<https://devfeed.tech/tags/backends.md>), [caching](<https://devfeed.tech/tags/caching.md>), [docker](<https://devfeed.tech/tags/docker.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [guide](<https://devfeed.tech/tags/guide.md>), [setup](<https://devfeed.tech/tags/setup.md>), [yaml](<https://devfeed.tech/tags/yaml.md>)

### AI overview

The right GitHub Actions cache depends on which work a build repeats: setup actions cache package stores, BuildKit handles Docker layers, and tools such as ccache, Bazel, Gradle, Turborepo, and Nx offer native caches for compile or task outputs. `actions/cache` lets workflows save named paths under exact keys, with prefix-based `restore-keys` as fallbacks; keys should include OS and relevant matrix dimensions and hash lockfiles rather than source trees. Its archives can be slow for large caches, while native per-artifact backends avoid restoring whole directories but require storage and eviction management. Self-hosted runners retain local disk state but require cleanup and increase the risk of jobs consuming one another's leftovers. The guide recommends matching the cache to the slow work, measuring restore time against work saved, and avoiding secrets, `.git`, high-churn data, and outputs whose keys omit OS or toolchain differences. It also describes co-located cache services as a faster option; the cited speed benchmarks and Avrea-specific features are vendor claims in the supplied text.

### Source excerpt

Every way to cache in GitHub Actions: actions/cache, setup actions, Docker layers, tool caches, and faster backends, with YAML for each.

## Introducing Avrea SBOM

DevFeed: [Introducing Avrea SBOM](<https://devfeed.tech/articles/introducing-avrea-sbom-122863.md>)

Original publisher: [Read original article](<https://avrea.com/blog/software-bill-of-materials>)

Author: Jeppe Tiensuu

Published: 2026-08-27T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [software composition analysis](<https://devfeed.tech/topics/software-composition-analysis.md>), [cyber resilience act](<https://devfeed.tech/topics/cyber-resilience-act.md>), [spdx](<https://devfeed.tech/topics/spdx.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [act](<https://devfeed.tech/tags/act.md>), [actively-exploited](<https://devfeed.tech/tags/actively-exploited.md>), [also](<https://devfeed.tech/tags/also.md>), [api](<https://devfeed.tech/tags/api.md>), [app](<https://devfeed.tech/tags/app.md>), [archive](<https://devfeed.tech/tags/archive.md>), [automatic](<https://devfeed.tech/tags/automatic.md>), [base](<https://devfeed.tech/tags/base.md>), [base-images](<https://devfeed.tech/tags/base-images.md>), [bill](<https://devfeed.tech/tags/bill.md>), [commit](<https://devfeed.tech/tags/commit.md>), [dependency](<https://devfeed.tech/tags/dependency.md>), [inventory](<https://devfeed.tech/tags/inventory.md>), [license](<https://devfeed.tech/tags/license.md>), [pinned](<https://devfeed.tech/tags/pinned.md>), [reports](<https://devfeed.tech/tags/reports.md>), [repositories](<https://devfeed.tech/tags/repositories.md>), [sbom](<https://devfeed.tech/tags/sbom.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

Avrea SBOM creates CycloneDX or SPDX dependency and license reports pinned to the analyzed commit, with per-commit history and automatic refreshes after relevant default-branch changes. It resolves dependencies across Python, Go, npm, Rust, Ruby, PHP, Maven, NuGet, container base images, and vendored source. Reports include production dependencies by default, plus packages shared with development or of undetermined scope; development-only dependencies can be included through the API. License gaps are reported explicitly rather than guessed, and reports omit timestamps and run IDs to make dependency changes easier to diff. Users can request a scan from a repository's SBOM tab. The excerpt describes the feature as live in preview and notes that older snapshots may require a new scan.

### Source excerpt

Avrea keeps a commit-pinned dependency and license inventory for connected repositories, with automatic refreshes and downloadable SBOM reports.

## How to Speed Up GitHub Actions

DevFeed: [How to Speed Up GitHub Actions](<https://devfeed.tech/articles/how-to-speed-up-github-actions-122864.md>)

Original publisher: [Read original article](<https://avrea.com/blog/speed-up-github-actions>)

Author: Tatu Mäkijärvi

Published: 2026-08-27T00:00:00Z

Content type: tutorial

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [amd-epyc](<https://devfeed.tech/tags/amd-epyc.md>), [builds](<https://devfeed.tech/tags/builds.md>), [cache](<https://devfeed.tech/tags/cache.md>), [concurrency](<https://devfeed.tech/tags/concurrency.md>), [faster](<https://devfeed.tech/tags/faster.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [matrix](<https://devfeed.tech/tags/matrix.md>), [sharding](<https://devfeed.tech/tags/sharding.md>), [speed](<https://devfeed.tech/tags/speed.md>)

### AI overview

Speed up GitHub Actions by finding which merge-blocking jobs and steps determine the wait, then fix the largest bottleneck and measure again. Repeated downloads may benefit from setup-action or tool caches, but a cache restore alone does not save work if tools reprocess files; branch-scoped cache writes and canceled runs can also leave later jobs cold. Concurrency groups can cancel superseded pull request runs and save runner minutes, while matrix sharding can cut test wall time but repeats per-job setup and does not reduce runner minutes. Change detection can skip unrelated work, though incorrect base commits or dependency graphs can silently omit tests, so the excerpt recommends comparing scoped and full runs before switching. Faster runners help CPU- or I/O-bound jobs, but more cores do not speed serial work. The excerpt also covers Docker layer cache configuration and warns that trigger filters can leave required checks unreported. Its benchmark and provider comparisons are presented as RunsOn/Avrea results, not universal performance guarantees.

### Source excerpt

Why GitHub Actions builds are slow, and what actually fixes it: cache hits, concurrency groups, matrix sharding, and when faster runners pay off.

## One sign-in for your whole organization: Avrea SAML SSO

DevFeed: [One sign-in for your whole organization: Avrea SAML SSO](<https://devfeed.tech/articles/one-sign-in-for-your-whole-organization-avrea-saml-sso-122862.md>)

Original publisher: [Read original article](<https://avrea.com/blog/saml-sso>)

Author: Hannu Valtonen

Published: 2026-08-18T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Single sign-on (SSO)](<https://devfeed.tech/topics/sso.md>), [saml](<https://devfeed.tech/topics/saml.md>)

Tags: [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [dns](<https://devfeed.tech/tags/dns.md>), [domains](<https://devfeed.tech/tags/domains.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [entra-id](<https://devfeed.tech/tags/entra-id.md>), [identity-provider](<https://devfeed.tech/tags/identity-provider.md>), [just](<https://devfeed.tech/tags/just.md>), [okta](<https://devfeed.tech/tags/okta.md>), [provisioning](<https://devfeed.tech/tags/provisioning.md>), [saml](<https://devfeed.tech/tags/saml.md>), [sso](<https://devfeed.tech/tags/sso.md>)

### AI overview

Avrea Enterprise adds organization-level SAML SSO for CI/CD access, connecting to any SAML 2.0 identity provider that exports metadata and signs assertions, including Okta, Microsoft Entra ID, Google Workspace, Keycloak, and OneLogin. Members authenticate under their IdP's policies; Avrea stores no passwords. Optional just-in-time provisioning creates memberships with a chosen default role, and verified email domains can be required to use SSO, blocking direct GitHub or Google sign-in for those members. Setup uses IdP metadata, attribute mapping, DNS TXT domain verification, and a test sign-in before enforcement. SAML establishes identity while linked GitHub permissions continue to control repository access. The article says SSO is included with Avrea Enterprise.

### Source excerpt

SAML SSO is available with Avrea Enterprise. Put CI/CD access behind Okta, Entra ID, or any SAML 2.0 identity provider, with just-in-time provisioning, verified domains, and enforced sign-in.

## Static IPs for GitHub Actions without managing runners

DevFeed: [Static IPs for GitHub Actions without managing runners](<https://devfeed.tech/articles/static-ips-for-github-actions-without-managing-runners-122851.md>)

Original publisher: [Read original article](<https://avrea.com/blog/github-actions-static-ip>)

Author: Hannu Valtonen

Published: 2026-08-13T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>)

Tags: [apis](<https://devfeed.tech/tags/apis.md>), [autoscaling](<https://devfeed.tech/tags/autoscaling.md>), [build](<https://devfeed.tech/tags/build.md>), [cache](<https://devfeed.tech/tags/cache.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [charges](<https://devfeed.tech/tags/charges.md>), [ci](<https://devfeed.tech/tags/ci.md>), [database](<https://devfeed.tech/tags/database.md>), [datacenters](<https://devfeed.tech/tags/datacenters.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [egress](<https://devfeed.tech/tags/egress.md>), [ephemeral](<https://devfeed.tech/tags/ephemeral.md>), [fintech](<https://devfeed.tech/tags/fintech.md>), [firewall](<https://devfeed.tech/tags/firewall.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [github-hosted-runners](<https://devfeed.tech/tags/github-hosted-runners.md>), [ip](<https://devfeed.tech/tags/ip.md>), [ipv4](<https://devfeed.tech/tags/ipv4.md>), [nat](<https://devfeed.tech/tags/nat.md>)

### AI overview

Avrea routes GitHub Actions job traffic through organization-reserved static IPv4 addresses while jobs continue to run on clean, short-lived VMs. Organizations can reserve one or more addresses in supported datacenters and enable all or selected addresses in Network egress; legacy configurations retain one address. Add the routed addresses to service allow-lists. Avrea uses an encrypted tunnel to an Avrea-managed router for outbound IPv4, while Avrea service and colocated cache traffic stay on a direct path. Its per-repository egress firewall separately limits destinations, with required platform connections allowed. Compared with self-hosting, this avoids maintaining runner images, updates, scaling, capacity, and cleanup. Each reserved IP costs $100 per month, metered at $3.33 per started UTC day; disabling routing does not stop charges, so release the address to end them.

### Source excerpt

Use dedicated Avrea static IPs to allowlist CI traffic. Choose one address or several across supported datacenters, while jobs run in ephemeral VMs.

## The Fastest GitHub Actions Runners in 2026: CPU Benchmarks Are Half the Story

DevFeed: [The Fastest GitHub Actions Runners in 2026: CPU Benchmarks Are Half the Story](<https://devfeed.tech/articles/the-fastest-github-actions-runners-in-2026-cpu-benchmarks-are-half-the-story-122844.md>)

Original publisher: [Read original article](<https://avrea.com/blog/fastest-github-actions-runners>)

Author: Tatu Mäkijärvi

Published: 2026-07-10T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [cache-invalidation](<https://devfeed.tech/topics/cache-invalidation.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [about](<https://devfeed.tech/tags/about.md>), [actions](<https://devfeed.tech/tags/actions.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [build](<https://devfeed.tech/tags/build.md>), [cache](<https://devfeed.tech/tags/cache.md>), [compare](<https://devfeed.tech/tags/compare.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [half](<https://devfeed.tech/tags/half.md>), [latest](<https://devfeed.tech/tags/latest.md>), [results](<https://devfeed.tech/tags/results.md>), [runner](<https://devfeed.tech/tags/runner.md>), [story](<https://devfeed.tech/tags/story.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

In RunsOn's November 2, 2026 comparison of four-vCPU x64 runners, Avrea completed Rust, TypeScript, and Docker build steps in 7m03s versus GitHub's 18m34s. The ranking sums median build-step times with build caches disabled; queueing and runner overhead are separate, and Avrea is the publisher's own product. RunsOn's 4 GB cache test measured Avrea restore at 11 seconds versus GitHub at 46 seconds, but the test does not represent every cache payload or backend. The excerpt stresses that CPU scores and transfer tests are not whole-workflow speedups: queue, startup, cache hits, workload, architecture, and cost matter. It recommends comparing the same workflow under matched cold and warm conditions, realistic changes, and repeated runs while recording end-to-end time and cost per successful run.

### Source excerpt

Avrea leads RunsOn's 4-vCPU x64 build benchmark. Compare the latest build, CPU and cache results, with methods and runner sizes.

## Your CI is the new attack surface: lock down what it can reach

DevFeed: [Your CI is the new attack surface: lock down what it can reach](<https://devfeed.tech/articles/your-ci-is-the-new-attack-surface-lock-down-what-it-can-reach-122842.md>)

Original publisher: [Read original article](<https://avrea.com/blog/egress-firewall-defense-in-depth>)

Author: Hannu Valtonen

Published: 2026-07-02T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [axios supply chain attack](<https://devfeed.tech/topics/axios-supply-chain-attack.md>)

Tags: [attack-surface](<https://devfeed.tech/tags/attack-surface.md>), [ci](<https://devfeed.tech/tags/ci.md>), [egress](<https://devfeed.tech/tags/egress.md>), [exfiltration](<https://devfeed.tech/tags/exfiltration.md>), [statistics](<https://devfeed.tech/tags/statistics.md>)

### AI overview

Avrea's Enterprise egress firewall is designed to limit data exfiltration from compromised CI jobs. Rules can be set by organization or repository using CIDR or hostname, with protocol and port filters. Policies default to allow-by-default; deny-by-default restricts traffic to an allow list but preserves required connections to GitHub, Avrea cache services, `*.blob.core.windows.net`, and `packages.microsoft.com`. The Azure Blob exception covers all Azure storage accounts, so it does not eliminate every exfiltration path. Enforcement uses `nftables` on Linux, `pf` on macOS, and Avrea's DNS resolver; additional safeguards block access to cloud metadata, other VMs, and SMTP. Per-run summaries show destinations and bytes, total egress and protocols, firewall-blocked traffic, and denied DNS lookups. The article suggests observing traffic under allow-by-default before building a deny-by-default policy. It presents egress control as one defense layer, not a complete solution.

### Source excerpt

Avrea's per-repository egress firewall plus per-run statistics make data and secret exfiltration much harder, and give you a record of exactly where every build connected.

## Guide to GitHub Actions Runners: Hosted, Self-Hosted, and Third-Party Options

DevFeed: [Guide to GitHub Actions Runners: Hosted, Self-Hosted, and Third-Party Options](<https://devfeed.tech/articles/guide-to-github-actions-runners-hosted-self-hosted-and-third-party-options-122850.md>)

Original publisher: [Read original article](<https://avrea.com/blog/github-actions-runners>)

Author: Tatu Mäkijärvi

Published: 2026-06-30T00:00:00Z

Content type: tutorial

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [Managed Services](<https://devfeed.tech/topics/managed-services.md>), [Self-hosted](<https://devfeed.tech/topics/self-hosted.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [GitHub](<https://devfeed.tech/topics/github.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2025](<https://devfeed.tech/tags/2025.md>), [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [about](<https://devfeed.tech/tags/about.md>), [account](<https://devfeed.tech/tags/account.md>), [actions](<https://devfeed.tech/tags/actions.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding-tools](<https://devfeed.tech/tags/ai-coding-tools.md>), [costs](<https://devfeed.tech/tags/costs.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [guide](<https://devfeed.tech/tags/guide.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [options](<https://devfeed.tech/tags/options.md>), [party](<https://devfeed.tech/tags/party.md>), [self-hosted](<https://devfeed.tech/tags/self-hosted.md>), [third-party](<https://devfeed.tech/tags/third-party.md>), [tradeoffs](<https://devfeed.tech/tags/tradeoffs.md>)

### AI overview

GitHub-hosted runners require little setup and offer GitHub's built-in logs and metrics, while self-hosted runners provide control over hardware, private-network access, and data location at the cost of fleet operations and added security risks. Third-party managed runners keep workflows largely intact through a `runs-on` label change while providers manage fleet, caching, scaling, and maintenance. The article recommends choosing by constraint: stay hosted for light builds, self-host when infrastructure control is necessary and a platform team can support it, or consider managed runners for slow builds without an internal fleet project. It suggests duplicating workflows and comparing warm-cache results before switching. Its cited benchmarks are workload-specific, and it notes that GitHub's proposed self-hosted minute fee was postponed and was not in effect as of June 2026.

### Source excerpt

Hosted, self-hosted, or third-party GitHub Actions runners? A practical breakdown of the tradeoffs, real costs, and how to choose, written for engineers.

## Avrea CLI: The full Avrea experience in your terminal, for humans and agents alike

DevFeed: [Avrea CLI: The full Avrea experience in your terminal, for humans and agents alike](<https://devfeed.tech/articles/avrea-cli-the-full-avrea-experience-in-your-terminal-for-humans-and-agents-alike-122866.md>)

Original publisher: [Read original article](<https://avrea.com/blog/the-full-avrea-experience-in-your-terminal>)

Author: Leo Lännenmäki

Published: 2026-06-22T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Terminal](<https://devfeed.tech/topics/terminal.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [ssh](<https://devfeed.tech/topics/ssh.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [avr](<https://devfeed.tech/tags/avr.md>), [cache](<https://devfeed.tech/tags/cache.md>), [cli](<https://devfeed.tech/tags/cli.md>), [client](<https://devfeed.tech/tags/client.md>), [command-line](<https://devfeed.tech/tags/command-line.md>), [live](<https://devfeed.tech/tags/live.md>), [manage](<https://devfeed.tech/tags/manage.md>), [run](<https://devfeed.tech/tags/run.md>), [running](<https://devfeed.tech/tags/running.md>), [ssh](<https://devfeed.tech/tags/ssh.md>), [terminal](<https://devfeed.tech/tags/terminal.md>), [triage](<https://devfeed.tech/tags/triage.md>), [vm](<https://devfeed.tech/tags/vm.md>), [watch](<https://devfeed.tech/tags/watch.md>), [workflows](<https://devfeed.tech/tags/workflows.md>)

### AI overview

Avrea's `avr` CLI brings run triage, job monitoring, VM access, workflow metrics, and cache/status views into the terminal. It scopes run listings from the Git remote; `run view --log-failed` groups failed steps by job, while run or job log commands can show full logs, follow live output, or filter by level and step. Job metrics show CPU and memory (optionally other sources), with five-second updates in watch mode; workflow views report run counts, timing percentiles, failure and flake rates over a default 30-day window, adjustable with `--since`. `job ssh` connects to the running job VM, and `--print-command` emits an SSH command for other tools. For automation, commands support field schemas and JSON projection, jq filtering, pipe-aware output, newline-delimited JSON watch events, and deterministic exit codes; the excerpt says exit code 4 indicates authentication is required and should be surfaced without retrying. The article presents these as ways for people and agents to inspect and debug Avrea workflows.

### Source excerpt

avr is Avrea's command-line client. Triage a failing run, watch it live, SSH onto the running VM, manage cache and workflows, all without leaving the terminal.

## M5 Max runners are here: macOS and ARM Linux CI on Apple Silicon

DevFeed: [M5 Max runners are here: macOS and ARM Linux CI on Apple Silicon](<https://devfeed.tech/articles/m5-max-runners-are-here-macos-and-arm-linux-ci-on-apple-silicon-122857.md>)

Original publisher: [Read original article](<https://avrea.com/blog/m5-max-runners-macos-arm-linux-ci-on-apple-silicon>)

Author: Hannu Valtonen

Published: 2026-06-09T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [macOS](<https://devfeed.tech/topics/macos.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Xcode](<https://devfeed.tech/topics/xcode.md>)

Tags: [actions](<https://devfeed.tech/tags/actions.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-silicon](<https://devfeed.tech/tags/apple-silicon.md>), [arm](<https://devfeed.tech/tags/arm.md>), [arm64](<https://devfeed.tech/tags/arm64.md>), [build](<https://devfeed.tech/tags/build.md>), [build-system](<https://devfeed.tech/tags/build-system.md>), [builds](<https://devfeed.tech/tags/builds.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [c](<https://devfeed.tech/tags/c.md>), [c-plus-plus](<https://devfeed.tech/tags/c-plus-plus.md>), [cache](<https://devfeed.tech/tags/cache.md>), [caching](<https://devfeed.tech/tags/caching.md>), [ci](<https://devfeed.tech/tags/ci.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [linux](<https://devfeed.tech/tags/linux.md>), [m5-max](<https://devfeed.tech/tags/m5-max.md>), [macos](<https://devfeed.tech/tags/macos.md>), [performance](<https://devfeed.tech/tags/performance.md>)

### AI overview

Avrea offers GitHub Actions runners on M5 Max Apple Silicon for macOS and ARM Linux, with caching intended to reduce repeated dependency downloads and builds. The article says its 8-core macOS runner costs $0.08 per minute, compared with GitHub's M2 Pro 5-core runner at $0.102 per minute (figures as of writing). It describes Xcode caching for Apple-platform workflows and ecosystem caches for ARM Linux, including Go, Rust, Node.js, Python, C/C++ and Zig when configured. Teams can generally try the runners by changing the workflow's `runs-on` label; the article recommends matching runner size to observed CPU and memory use. It names example labels for macOS and Ubuntu ARM jobs, while directing readers to current labels for available sizes.

### Source excerpt

Avrea now supports GitHub Actions workloads on macOS and ARM Linux powered by the latest M5 Max Apple Silicon. Less money for much more performance.

## Avrea is SOC 2 Type 2 attested

DevFeed: [Avrea is SOC 2 Type 2 attested](<https://devfeed.tech/articles/avrea-is-soc-2-type-2-attested-122837.md>)

Original publisher: [Read original article](<https://avrea.com/blog/avrea-is-soc-2-type-2-attested>)

Author: Jeppe Tiensuu

Published: 2026-06-03T00:00:00Z

Content type: release

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [soc 2](<https://devfeed.tech/topics/soc-2.md>)

Tags: [attestation](<https://devfeed.tech/tags/attestation.md>), [audit](<https://devfeed.tech/tags/audit.md>), [audits](<https://devfeed.tech/tags/audits.md>), [automated](<https://devfeed.tech/tags/automated.md>), [availability](<https://devfeed.tech/tags/availability.md>), [ci](<https://devfeed.tech/tags/ci.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [iso-27001](<https://devfeed.tech/tags/iso-27001.md>), [overlap](<https://devfeed.tech/tags/overlap.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [report](<https://devfeed.tech/tags/report.md>), [secrets](<https://devfeed.tech/tags/secrets.md>), [security](<https://devfeed.tech/tags/security.md>), [soc-2](<https://devfeed.tech/tags/soc-2.md>)

### AI overview

Avrea reports SOC 2 Type 2 attestation covering Security, Availability, and Confidentiality. Its initial report covers 9 February to 9 May 2026 and will be renewed annually; the full report is available under NDA. Avrea says it designed audit controls before beginning and ran the SOC 2 and ISO 27001 audits in parallel, citing roughly two-thirds overlap between the frameworks and automated evidence reports to support ongoing compliance. The attestation may help teams evaluating Avrea as a CI vendor, given the platform handles source code, secrets, caches, and build artifacts.

### Source excerpt

Avrea is SOC 2 Type II attested. Security has been part of the platform from day one, not something added later. Learn what the attestation covers and what it means for teams building with Avrea.

## CI/CD for the age of AI

DevFeed: [CI/CD for the age of AI](<https://devfeed.tech/articles/ci-cd-for-the-age-of-ai-122855.md>)

Original publisher: [Read original article](<https://avrea.com/blog/launching-avrea>)

Author: Hannu Valtonen

Published: 2026-05-26T00:00:00Z

Content type: news

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [CI/CD](<https://devfeed.tech/topics/cicd.md>), [Continuous Delivery (CD)](<https://devfeed.tech/topics/continuous-delivery.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>), [build times](<https://devfeed.tech/topics/build-times.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [Legacy Modernization](<https://devfeed.tech/topics/legacy-modernization.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [age-of-ai](<https://devfeed.tech/tags/age-of-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [avrea](<https://devfeed.tech/tags/avrea.md>), [build](<https://devfeed.tech/tags/build.md>), [cd](<https://devfeed.tech/tags/cd.md>), [ci](<https://devfeed.tech/tags/ci.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [development](<https://devfeed.tech/tags/development.md>), [faster](<https://devfeed.tech/tags/faster.md>), [funding](<https://devfeed.tech/tags/funding.md>), [funding-round](<https://devfeed.tech/tags/funding-round.md>), [more](<https://devfeed.tech/tags/more.md>), [native](<https://devfeed.tech/tags/native.md>), [quality](<https://devfeed.tech/tags/quality.md>), [ship](<https://devfeed.tech/tags/ship.md>), [software](<https://devfeed.tech/tags/software.md>), [software-development](<https://devfeed.tech/tags/software-development.md>), [teams](<https://devfeed.tech/tags/teams.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

Avrea says AI-generated code has made software delivery the bottleneck: traditional CI/CD pipelines may not scale as code, tests and validations multiply. Its platform is designed to work with existing CI/CD workflows after a one-line code change and to let AI agents use the system. Operating inside the CI environment, it can inspect logs, cache behavior, tests, dependencies and infrastructure bottlenecks. The company says it is building automated diagnosis and proposed fixes, including pull requests, dependency upgrades validated against CI, tooling modernization, test improvements and evolving documentation; proposed changes are validated against the full pipeline before developer review. Avrea announced $4.7 million in total Pre-Seed funding led by Earlybird. Eligible organizations can try it with monthly credits equivalent to 3,000 minutes on 2-vCPU Linux x64 runners, subject to eligibility and usage limits.

### Source excerpt

Introducing Avrea and our funding round as we build CI/CD for AI-native software development, helping teams ship faster, more often, and with higher quality.

## How we made GitHub Actions cache up to 6x faster

DevFeed: [How we made GitHub Actions cache up to 6x faster](<https://devfeed.tech/articles/how-we-made-github-actions-cache-up-to-6-faster-122852.md>)

Original publisher: [Read original article](<https://avrea.com/blog/how-we-made-github-actions-cache-up-to-6x-faster>)

Author: Leo Lännenmäki

Published: 2026-05-21T00:00:00Z

Content type: article

Language: en

Sources: [Avrea Blog](<https://devfeed.tech/sources/avrea-blog.md>)

Topics: [Cache](<https://devfeed.tech/topics/cache.md>), [GitHub Actions](<https://devfeed.tech/topics/github-actions.md>), [Architecture & optimization](<https://devfeed.tech/topics/architecture-optimization.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>), [interface implementation](<https://devfeed.tech/topics/interface-implementation.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>)

Tags: [1-5](<https://devfeed.tech/tags/1-5.md>), [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [actions](<https://devfeed.tech/tags/actions.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [api](<https://devfeed.tech/tags/api.md>), [azure](<https://devfeed.tech/tags/azure.md>), [backend](<https://devfeed.tech/tags/backend.md>), [cache](<https://devfeed.tech/tags/cache.md>), [client](<https://devfeed.tech/tags/client.md>), [faster](<https://devfeed.tech/tags/faster.md>), [github](<https://devfeed.tech/tags/github.md>), [github-actions](<https://devfeed.tech/tags/github-actions.md>), [restore](<https://devfeed.tech/tags/restore.md>), [takes](<https://devfeed.tech/tags/takes.md>), [up](<https://devfeed.tech/tags/up.md>), [variance](<https://devfeed.tech/tags/variance.md>), [with](<https://devfeed.tech/tags/with.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

Avrea reports that its GitHub Actions cache restores are up to 6x faster than GitHub-hosted restores, with less variation on larger entries. The comparison measures end-to-end `actions/cache/restore` step duration, including transfer, decompression, and file I/O; each reported value is the median of 50 runs. Avrea achieves compatibility without changing the workflow or cache client: a local Go proxy implements the required Twirp and Azure blob operations, verifies GitHub-signed runtime tokens, and routes other traffic to GitHub. Co-located API and storage reduce network round trips, while faster runner machines help with transfer and processing. Small caches see smaller gains because request overhead dominates. The approach must preserve GitHub's branch and repository access rules, and the post notes `actions/cache` is less suited to caching many small entries.

### Source excerpt

On GitHub Actions, a 1 GB actions/cache restore takes 14 seconds. On Avrea, the same workflow, same client: 2.9 seconds. Up to 6x faster across cache sizes from 1 KB to 1 GB, with 3x tighter variance.

[Next page](<https://devfeed.tech/sources/avrea-blog.md?cursor=WyIyMDI2LTA1LTIxVDAwOjAwOjAwKzAwOjAwIiwgIjdjZTNkM2QxLWIzNDQtNDk1Yy1hMjljLWVmOWVmZWI4MWI1MSJd>)