Every metric keeps being collected and stays available for alarms. Unchecked metrics are only hidden from this object's dashboard charts.
This token is shown once. Save it securely on the monitored server.
AWS resources are monitored through this account. Credentials are stored encrypted on the monitoring server and are never shown again.
Choose where alarm events should be delivered.
Paste the health manifest JSON of a service. Its checks are called one after another from the monitoring server, and are also polled automatically on a schedule (default every 60s, override with a top-level "pollSeconds", or set "enabled": false to pause polling). Add "enabled": false to an individual check to exclude just that one from the schedule. If the service's health endpoint needs a credential, add "auth": { "header": "x-api-key", "value": "…" } inside "service" — it is stored encrypted and never shown again, so editing a manifest without re-entering it keeps the saved value, and "auth": null clears it.
Selected metrics are added as chart widgets. Metrics with different units become separate charts.
Search for a monitored resource, then pick the metrics to chart.