EvoHub Docs Sign in
DeutschDE
EnglishDeutsch
Diese Seite liegt noch nicht auf Deutsch vor und wird auf Englisch angezeigt.

Uptime

Heartbeat monitors

Mit KI verwenden
Als Markdown anzeigenDiese Seite als reiner Text, zum Einfügen in ein KI-Tool In Claude öffnenClaude Fragen zu dieser Seite stellen In ChatGPT öffnenChatGPT Fragen zu dieser Seite stellen
Mit Cursor / VS Code / Claude verbindenDiese Dokumentation aus Ihrem KI-Tool durchsuchen und lesen (MCP-Server)

URL des MCP-Servers

https://docs-dev.evohub.io/mcp

Zu Cursor hinzufügenZu VS Code hinzufügen

Claude Code

claude mcp add --transport http evohub-docs-docs https://docs-dev.evohub.io/mcp

Claude (claude.ai und Claude Desktop): Einstellungen → Konnektoren → Benutzerdefinierten Konnektor hinzufügen und die URL oben einfügen.

Claude Desktop — claude_desktop_config.json

{
  "mcpServers": {
    "evohub-docs-docs": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://docs-dev.evohub.io/mcp"
      ]
    }
  }
}

Cursor — ~/.cursor/mcp.json

{
  "mcpServers": {
    "evohub-docs-docs": {
      "url": "https://docs-dev.evohub.io/mcp"
    }
  }
}

VS Code — .vscode/mcp.json

{
  "servers": {
    "evohub-docs-docs": {
      "type": "http",
      "url": "https://docs-dev.evohub.io/mcp"
    }
  }
}

llms.txtllms-full.txt

A heartbeat monitor watches something that runs on a schedule, such as a cron job, a nightly backup or a queue worker. Instead of EvoHub checking a target, your job calls a unique URL each time it runs. If a call does not arrive in time, EvoHub marks the monitor down and alerts you, the way it would for any other monitor.

When to use one

Use a heartbeat monitor when the thing you care about has no address to check, or when "it ran" is what matters:

  • scheduled jobs and cron tasks
  • backups and data exports
  • background workers that should check in regularly
  • anything that fails silently by simply not running

Create a heartbeat monitor

Start a new monitor

In Uptime, click New Monitor, enter a Name, and set Monitor Type to Heartbeat (cron / job).

Set the expected ping interval

Under Expected ping interval, enter how often your job runs, as a number and a unit: minutes, hours or days. For example, 1 days for a nightly job or 3 hours for one that runs every three hours.

Set the grace period

Under Grace period, enter how late a ping may be before the monitor goes down, in minutes or hours. Allow for how long your job takes and how much its start time varies.

Choose alerting

Optionally pick an Escalation Policy under Alerting to page your team through On-Call. You can also add Tags.

Create

Click Create Monitor. The monitor's page shows its ping URL.

Ping the URL

The monitor's page has a Heartbeat ping URL card with the address to call and a Copy button. The URL looks like https://evohub.io/ping/<token>; always copy the exact one from the console.

Call it at the end of your job, after the work has succeeded. GET, POST and HEAD all work, and no API key or other credential is needed.

cron

# Runs at 02:00 every day, and pings only if the backup succeeds
0 2 * * * /usr/local/bin/backup.sh && curl -fsS https://evohub.io/ping/<token>

shell script

#!/usr/bin/env bash
set -euo pipefail
./run-export.sh
curl -fsS --retry 3 https://evohub.io/ping/<token>

A successful ping returns 200 with {"status":"ok"}. An unknown URL returns 404.

The card also shows Last ping, so you can confirm your job is reaching EvoHub.

Warning

The ping URL is the only credential. Anyone who has it can reset the monitor's timer. Keep it out of public repositories and logs.

When a heartbeat goes down

The monitor is down when the time since the last ping is longer than the expected interval plus the grace period.

For example, with an interval of 1 day and a grace period of 30 minutes, a job that last pinged at 02:05 on Monday is marked down if no ping arrives by 02:35 on Tuesday.

  • A new monitor that has never been pinged gets one full interval plus grace, counted from when you created it, before it can go down.
  • When the monitor goes down, EvoHub opens an incident with the reason, for example "no heartbeat received for 25h0m0s", and alerts your notification channels and On-Call, just like a failed HTTP check.
  • While it stays down, you are alerted again on the monitor's Re-alert while still down schedule.
  • The next ping brings the monitor back up and closes the incident.

How heartbeats differ from other monitors

  • There is no target, timeout, assertion, slow threshold or SLA target.
  • The monitor page records state changes rather than every ping, so Recent Checks shows when it went down and came back.
  • Heartbeat monitors are not metered for billing.

Pause or silence a heartbeat

Pause on the monitor's page stops it from being evaluated, for example while a job is switched off. To mute alerts for a planned window instead, use a silence window. See Uptime alerts and silence windows.

Zuletzt aktualisiert am