> Index of every page on this site: [llms.txt](https://docs.a-line.cloud/llms.txt)

# Run GitHub Actions jobs on A-Line.cloud

> Connect GitHub, pick a runner size with a runs-on label and cap what your CI spends.

**Goal**: GitHub Actions jobs that run on A-Line.cloud runners, billed per
vCPU-minute, with a spend cap you control.

## Before you start

- A GitHub organization or account where you can install a GitHub App.
- Credit on your balance. Jobs are paid as they run; see
  [Billing](/billing).

## Connect GitHub

<Steps>
<Step>

Open **Actions** in the sidebar and go to the **Setup** tab.

</Step>
<Step>

Select **Enable GitHub Actions runners** (or **Add GitHub account** for
another account). This installs the A-Line.cloud Actions app on GitHub.
Pick the repositories it may serve.

</Step>
<Step>

Back on the **Setup** tab, the account is listed. Jobs of these accounts
with an `a-line` label run on A-Line.cloud.

</Step>
</Steps>

**Run jobs from fork pull requests** is off by default: fork code would run
on your runners and be billed to you. If you turn it on, set **Require
approval for all outside collaborators** in the repository or organization
Actions settings on GitHub.

## Pick a runner

<Steps>
<Step>

On the **Setup** tab, use the runner picker: choose the vCPU and memory,
the **Image** and the disk.

</Step>
<Step>

Select **Copy label** and paste the **Runner label** into your workflow:

</Step>
</Steps>

```yaml
jobs:
  test:
    runs-on: a-line-4vcpu
    timeout-minutes: 30
    steps:
      - uses: actions/checkout@v4
      - run: make test
```

The label reads `a-line-<N>vcpu`, with optional parts for memory, disk, OS,
the **Full** image and ARM, in this order:
`a-line-<N>vcpu[-<M>gb][-<D>gb][-<os>-<version>][-full][-arm]`.

- Sizes go from 2 to 160 vCPU, with 4 GB of memory per vCPU by default.
- The **Full** image is close to GitHub's own runner image; use it when a
  job expects tools that the **Standard** image does not have.
- 50 GB of disk is included. More is billed per GB-hour.

### Find your label

<RunnerPicker>

For example `a-line-8vcpu` is 8 vCPU with 32 GB of memory, and
`a-line-8vcpu-200gb-ubuntu-24.04` adds 200 GB of local disk and pins Ubuntu
24.04.

</RunnerPicker>

## Cap the spend

On the **Limits** tab:

- **Job duration limit (minutes)**: jobs running longer are stopped. Leave
  it empty for no limit.
- **Monthly spend cap (EUR)**: covers Actions and build compute. Only
  organization admins can change it. The console warns you when the cap is
  near and when it is reached.

Set `timeout-minutes` on each job too. GitHub's default is 360 minutes.

## Verify

Push a commit that runs the workflow. On GitHub, the job shows it was picked
up by an A-Line.cloud runner. In the console, the job appears in the history
on the **Overview** tab. Time spent waiting in the queue is not billed.

## Troubleshooting

- **The job waits forever for a runner**: the label is wrong, or the
  repository is not selected in the GitHub App. Copy the label from the
  picker again and check the app's repository access on GitHub.
- **The Setup tab shows No webhook received yet**: GitHub has not sent an
  event since the app was installed. Start a workflow run.
- **The account shows Suspended on GitHub**: the app was suspended in your
  GitHub settings. Unsuspend it there.
- **Jobs wait with Cap reached**: new runners wait until the **Monthly spend
  cap (EUR)** is raised or the month ends.
- **You need more runners at once or a larger size**: select **Request
  higher limits**; see [Limits](/actions/limits).

## Next steps

- [Runner sizing](/actions/runner-sizing): size suggestions from measured
  runs.
- [Actions insights](/actions/insights): flaky jobs and spend forecast.
- [Long-running jobs](/actions/long-running-jobs): jobs that run for hours.
