What is Monitoring?
Monitoring records metrics of your machines and other resources, alarms on them, keeps the logs your apps and load balancers ship, checks your URLs from outside, and publishes status pages.
Monitoring needs nothing installed on your machines. The platform records CPU, disk and network of every running compute instance and Managed PostgreSQL or Document Database node, and its own metrics of site-to-site VPN connections, Kubernetes clusters, CDN distributions and app deployments. Your own code can publish custom metrics next to them. An alarm watches one metric and mails you or calls a webhook when it crosses a threshold. You find Monitoring under Monitoring in the Deutsche Cloud Console; the charts of a resource are also on its own page.
When to use it
- You want to know when a machine is busy, a disk fills or a tunnel goes down, without running a monitoring stack of your own.
- Your application should publish its own numbers (orders, queue depth, latency) through the API, OpenTelemetry or StatsD, and alarm on them.
- You want to read what a deployed app printed, or the requests an application load balancer or a CDN distribution served, without opening a shell.
- A URL, on Deutsche Cloud or anywhere else, should be requested on a schedule and someone told when it stops answering.
- Your users should see a public page that says whether your service works, with the incidents you post.
Concepts
| Term | Meaning |
|---|---|
| Metric | One measurement of one resource over time, such as cpu_percent. See Metrics. |
| Custom metric | A series your own code publishes: a namespace, a metric name and dimensions. See Publish custom metrics. |
| Alarm | A rule on one metric: a statistic over a period, compared with a threshold. Its state is OK, ALARM or INSUFFICIENT_DATA. See How alarms work. |
| Log line | One line an app's container, a schedule, a function, an event rule, a load balancer or a CDN distribution shipped. See Search logs. |
| Uptime check | A URL requested on a schedule from every location the platform checks from. See How uptime checks work. |
| Status page | A public page at {slug}.status.<domain> with components, incidents and subscribers. See Create a status page. |
Pricing
Metrics of your resources, alarms, uptime checks and status pages are free. The log events and the custom metric points your organization keeps are metered by the GB-month, measured daily; where they are not charged yet, the cost shows as "metered, not charged yet" and is not on your invoice.
Interfaces
- The Deutsche Cloud Console, under Monitoring: Alarms, Custom metrics, Uptime, Status pages and Logs.
- The CLI:
dcloud monitoring …; see the CLI Command Reference. - The HTTP API under
/monitoring/api; see the API Reference. - Terraform:
dcloud_monitoring_alarm,dcloud_uptime_check,dcloud_status_pageanddcloud_status_page_subscriber.
Next steps
- Get started: publish a metric, alarm on it, check a URL, and clean up.
- Create an alarm
- Create an uptime check