GitHub Runner (non-dockerized)

Spin up an on-demand self-hosted GitHub action runner on any Unix like operating system.

Environment Variables

• ARTIFACT_VERSION - Runner version. Uses the latest version from actions/runner.
• ACTIONS_DIR - Directory where the runner has to be downloaded and configured. Uses the current working directory.
• GIT_TOKEN - Required for authentication to add runners.
• GIT_OWNER - GitHub account username [OR] organization name.
• GIT_REPOSITORY - Repository name (required to create runners dedicated to a particular repo)
• RUNNER_GROUP - Runner group. Uses default
• RUNNER_NAME - Runner name. Uses a random instance ID.
• WORK_DIR - Work directory. Uses _work
• LABELS - Runner labels (comma separated). Uses "docker-node,${os_name}-${architecture}"
• REUSE_EXISTING - Re-use existing configuration. Defaults to false

Note

REUSE_EXISTING flag can be useful when a container restarts due to an issue or when a container is reused after being terminated without shutting down gracefully.

More info

Following files/directories are created (commonly across macOS, Linux and Windows runners) only when the runner has been configured

• _work
• _diag
• .runner
• .credentials
• .credentials_rsaparams

So, a simple check on one or more of these files' presence should confirm if the runner has been configured already

Note: Warnings like the ones below are common, and GitHub typically reconnects the runner automatically.

A session for this runner already exists.

Runner connect error: The actions runner i-058175xh7908r2u46 already has an active session.. Retrying until reconnected.

Warning

Using this image without the env var GIT_REPOSITORY will create an organization level runner.
Using self-hosted runners in public repositories pose some considerable security threats.

• #self-hosted-runner-security
• #restricting-the-use-of-self-hosted-runners
• #configuring-required-approval-for-workflows-from-public-forks

Env vars for notifications

This project supports ntfy and telegram bot for startup/shutdown notifications.

NTFY

Choose ntfy setup instructions with basic OR authentication abilities

• NTFY_URL - Ntfy endpoint for notifications.
• NTFY_TOPIC - Topic to which the notifications have to be sent.
• NTFY_USERNAME - Ntfy username for authentication (if topic is protected)
• NTFY_PASSWORD - Ntfy password for authentication (if topic is protected)

Telegram

Steps for telegram bot configuration

1. Use BotFather to create a telegram bot token
2. Send a test message to the Telegram bot you created
3. Use the URL https://api.telegram.org/bot{token}/getUpdates to get the Chat ID
   • You can also use Thread ID to send notifications to a particular thread within a group

export TELEGRAM_BOT_TOKEN="your-bot-token"
export CHAT_ID=$(curl -s "https://api.telegram.org/bot${TELEGRAM_BOT_TOKEN}/getUpdates" | jq -r '.result[0].message.chat.id')

• TELEGRAM_BOT_TOKEN - Telegram Bot token
• TELEGRAM_CHAT_ID - Chat ID to which the notifications have to be sent.
• THREAD_ID - Optional thread ID to send notifications to a specific thread.

Note: To send notifications to threads, the bot should be added to a group with Topics enabled.
Send a message to the bot in a group thread

export THREAD_ID=$(curl -s "https://api.telegram.org/bot${TELEGRAM_BOT_TOKEN}/getUpdates" | jq -r '.result[0]|.update_id')

License & copyright

© Vignesh Rao

Licensed under the MIT License