Self-Hosted Github Actions Runners

How Going In-House Can Slash Your Build Times

The Drawback of GitHub-Hosted Runners

Ever wondered why your iOS builds take an eternity on GitHub Actions runners? You’re not alone. Here are some possible culprits:

  1. GitHub Action (GHA) runners primarily use Intel-based chips, which may not be as efficient as newer architectures.
  2. If you are not using caches, every new action has to go through multiple steps:
    • Container creation
    • Setup
    • Dependency download and installation
    • iOS simulator setup
    • Tests execution

Case Study: A Minimal Xcode Project

As an example, I’ve created a new Xcode project, with just 2 files inside, and only 1 unit test:

import XCTest
@testable import TestCI

final class TestCITests: XCTestCase {
    func testExample() throws {
        XCTAssertEqual(MyVC().hey, "Hey")
        XCTAssertEqual(ContentView().myInt, 0)

Then, I created this fastlane lane to run the tests:

platform :ios do
  lane :tests do
      scheme: "TestCI"

Finally, I configured a simple GitHub Actions workflow as follows:

name: iOS starter workflow

    branches: [ "main" ]
    branches: [ "main" ]

    name: Build and Test
    runs-on: macos-latest

      - name: Checkout
        uses: actions/checkout@v3
      - name: Bundle install
        run: |
          bundle install
      - name: Test
        run: |
          bundle exec fastlane tests



The GitHub-hosted runner took a staggering 5 minutes and 9 seconds (or 309 seconds) to complete this simple test.

A Faster Alternative: Self-Hosted Runners

Enter self-hosted runners. If you have an M1/M2 machine, here’s how you can set it up with minimal changes:

Fastlane Lane Configuration:

lane :tests do
    scheme: "TestCI",
    device: "iPhone 8", # Use always the same device 
    reset_simulator: false, # Avoid resetting it after the tests
    skip_detect_devices: true # Skip the devices detection

GitHub Actions Workflow:

  • Change: runs-on: macos-latest
  • To: runs-on: self-hosted

Quick Self-Hosted Setup Guide

This step was actually easier than I thought.

  1. Navigate to the Repository’s Settings on Github (you’d need admin rights)
  2. Go to Actions -> Runners
  3. Tap on Add Runner or New self-hosted runner
  4. Choose your OS and architecture (ARM64 for M-chips)
  5. Follow the steps on that screen to download and configure the self-hosted runner on your machine
  6. Finally, run your runner with ./

That’s it, now your runner will start listening for new jobs.



After this setup, my self-hosted runner executed the same test in a blazing 14 seconds, an almost 95% improvement!

Note: My machine had pre-downloaded dependencies (fastlane) and an open iPhone 8 simulator, which contributed to the speed.

Pros / Cons of going self-hosted

While the benefits are tempting, self-hosting is not a one-size-fits-all solution.


  • Faster Builds: A self-hosted runner is generally persistent, allowing you to retain build caches, derived data, and dependencies between runs. This can dramatically reduce build times.
  • Customization: You can tailor the environment to your exact needs, including installing specific software or tools that are critical to your iOS development workflow.
  • Performance: Tailoring hardware to your requirements could lead to faster build and test times, boosting productivity.
  • Resource Control: Being on-premise means you control the resources, ensuring they are used optimally, without other tasks affecting them.
  • Cost: It could be more cost-effective in the long run if you already own powerful hardware.
  • Security: All code and data remain within your local network, which may be a security requirement for some of your US-based clients.
  • Debugging: Easier to debug any issues when you have direct access to the environment where your code is being built and tested.


  • Maintenance: You’re responsible for updates, security patches, and generally keeping the machine running smoothly
  • Upfront Costs: There could be significant initial costs for powerful hardware, software licenses, and perhaps even cooling solutions.
  • Electricity Costs: Running a machine 24/7 can add to your utility bills.
  • Dependency: Your CI/CD pipeline would depend on the availability of your local machine. Power outages, hardware failures, or network issues could disrupt your workflow.
  • Scalability: As your projects grow, you might need to invest in more hardware or distribute tasks across multiple machines manually.
  • Remote Access: If you travel more (one of your goals), you’d need a reliable way to remotely access and manage your self-hosted runner.
  • Cache Pollution: If not carefully managed, caches can become polluted with outdated or unused data, leading to potential inconsistencies and bugs.

Final Thoughts: Is Self-Hosting for You?

With these considerations in mind, is it time for you to switch to a self-hosted runner for your iOS builds? The decision, as always, depends on your specific needs and constraints.

Self-Hosted Github Actions Runners |
Tags: tools testing
Share: Twitter LinkedIn