uptime-monitor

What This Skill Does

The uptime-monitor is a silent, persistent sentinel designed to track the health of your OpenClaw agent or gateway infrastructure. It operates on a strict 5-minute interval via cron, providing automated heartbeat verification without requiring manual intervention. The core functionality centers on state management: if the system is detected as live, it increments a rolling streak counter stored in workspace/uptime/streak.json. Upon achieving a 168-hour (7-day) continuous alive streak, the skill automatically generates an uptime.json manifest to serve as a proof-of-uptime record. Conversely, if a ping fails, the skill writes a dead.json file containing the timestamp of the outage, triggering a reset of the streak counter. It is a completely self-contained, file-based monitoring solution that requires no external databases or network dependencies.

Installation

To integrate the uptime-monitor into your environment, use the OpenClaw Hub command: clawhub install openclaw/skills/skills/kunoiiv/uptime-monitor. Once installed, you must initialize the cron job to handle the heartbeat pings. Run: 📊 cron add uptime-5m '{"kind":"every","everyMs":300000}' '{"kind":"systemEvent","text":"UPTIME CHECK 👻"}' --sessionTarget main. This establishes the periodic triggers required for the logic to execute autonomously in the background.

Use Cases

This skill is perfect for infrastructure reliability tracking, automated developer tool monitoring, and verifying agent persistence. It is best used when you need a low-overhead, file-based audit trail for your OpenClaw instance, such as when deploying on a remote VPS where you need to know if the agent crashed during maintenance windows. It also serves as a diagnostic foundation for automated restart scripts or incident response logging.

Example Prompts

1. "OpenClaw, verify the current health status and tell me if my uptime streak is still active."
2. "Check the workspace/uptime folder and summarize the last known downtime event from dead.json."
3. "Manually trigger an UPTIME CHECK 👻 and verify if the streak counter incremented correctly."

Tips & Limitations

The uptime-monitor is a passive system. It does not send push notifications; you must check the files directly or use OpenClaw commands to read them. Ensure your disk has write permissions in the workspace directory for the auto-mkdir to function correctly. If the cron job is interrupted, the streak is paused rather than reset, preventing false positives during planned system reboots. Always check streak.json to verify that the last_ping timestamp aligns with your expected uptime expectations.