# Autopilot ## Overview {% embed url="" %} Most infrastructure problems don’t start as incidents. They start as: * a pod getting OOM Killed * disk quietly hitting 100% * a “we’ll fix it later” change And by the time you notice… It’s already expensive. **Stakpak Autopilot changes that.** It doesn’t just monitor. It runs 24/7, detects unexpected changes, fixes what’s safe, and only alert you when it actually matters. No dashboards to babysit.\ No alerts to ignore.\ No “we’ll look at it tomorrow.” ### Getting Started Takes Less Than a Min 1. [Install Stakpak](/docs/get-started/install-stakpak.md) 2. Run [/Init](/docs/how-it-works/init.md) (understand your apps and tech stack) 3. Start with the recommended schedules That’s it. ## Stakpak Autopilot Prerequisite Before running autopilot: * Docker must be installed and accessible to the current user * 2GB+ RAM is recommended for reliable autopilot + sandbox runs * Swap is strongly recommended on small Linux hosts * Linux user services may require linger to survive logout ## **Channels** You can connect Stakpak Autopilot to * Discord * Telegram * Slack ([slack yaml](https://github.com/stakpak/agent/blob/main/libs/gateway/src/channels/slack-manifest.yaml)) ## Setting Up Stakpak Autopilot
All you have to do is open Stakpak and ask it to monitor whatever you want, and it will help you set up the channels and everything {% hint style="success" %} Did you know you can either run Stakpak autopilot locally or on a remote server {% endhint %} ### Use Cases
You can use Stakpak Autopilot to monitor your infrastructure 24/7, automatically fix what’s safe, and alert you only when it actually needs you: * Monitor a fleet of OpenClaw deployments (gateway health, agents, connectivity) * Keep nagging you until you actually set up backups for your database * Detect and fix pods getting OOMKilled before users notice * Restart stuck services or failing jobs automatically * Detect expired credentials / broken integrations and fix or escalate * Spot duplicate or unused resources quietly burning money * Alert you when something *actually* needs a decision (not noise) * Investigate failed deployments and suggest/apply fixes * Continuously check that critical services are recoverable (not just “running”) ### **Configuration** * `~/.stakpak/config.toml`: [Profiles](/docs/how-it-works/profiles.md) behavior (`model`, `allowed_tools`, `auto_approve`, `system_prompt`, `max_turns`, provider credentials) * `~/.stakpak/autopilot.toml`: runtime wiring (`schedules`, `channels`, service/server settings) Use `profile = "name"` on schedules/channels and keep behavior inside profile definitions. ``` # schedule profile stakpak autopilot schedule add health --cron '*/5 * * * *' --prompt 'Check health' --profile monitoring # channel profile stakpak autopilot channel add slack --bot-token "$SLACK_BOT_TOKEN" --app-token "$SLACK_APP_TOKEN" --profile ops ``` Full setup guide: [cli/README.md](https://github.com/stakpak/agent/blob/main/cli/README.md) ### Commands ``` stakpak autopilot up # Start autopilot and install as system service (runs setup on first use) stakpak autopilot down # Stop autopilot and remove system service stakpak autopilot status # Show health, uptime, schedule/channel metadata, and recent activity stakpak autopilot logs # Stream autopilot logs stakpak autopilot restart # Restart autopilot (reload config) stakpak autopilot schedule # Manage scheduled tasks stakpak autopilot channel # Manage messaging channels (Slack, Telegram, Discord) stakpak autopilot doctor # Run preflight checks for autopilot setup/runtime ``` #### Alias ``` stakpak up # alias for: stakpak autopilot up stakpak down # alias for: stakpak autopilot down ``` ### References * [Install Stakpak](/docs/get-started/install-stakpak.md) * [/Init](/docs/how-it-works/init.md) * [Stakpak/Agent Readme](https://github.com/stakpak/agent) * [cli/README.md](https://github.com/stakpak/agent/blob/main/cli/README.md) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://stakpak.gitbook.io/docs/how-it-works/autopilot.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.