Monitoring and Alerts
Constella collects metrics from connected clouds and virtual machines, displays them, and alerts you when something crosses a threshold.
Look at metrics
- Monitoring › Overview summarizes usage and alerts across the platform.
- Metric › Dashboard holds charts you arrange into dashboards, for example a dashboard per application or per team.
- Metric › Explorer is for ad-hoc questions. Pick a resource type and a metric (CPU, memory, disk or network), filter and group the results, and save the chart to a dashboard when it is useful.
Metrics for cloud resources come from each provider's monitoring service. How quickly they appear and how fine-grained they are depends on the provider.
Alert when something goes wrong
- Open Alerting › Policies and click Create.
- Choose what to watch: the resource type, the metric, and optionally which resources (for example by project or tag).
- Set the condition, for example CPU above 90% for 5 minutes, and the alert level.
- Choose who is notified and how (see below).
Triggered alerts appear under Alerting › Resources (resources that are currently alerting) and in the alert history.
Notifications
Notifications reach people through channels and recipients.
| Menu | Purpose |
|---|---|
| Channels | How notifications leave Constella: email (SMTP), SMS, and the DingTalk, Feishu and WeCom apps |
| Recipients | People and their contact details. A contact is used only after it has been verified. |
| Bot | Group chat bots and webhooks: DingTalk, Feishu and WeCom bots, or any HTTP webhook, for example to reach a chat tool or an incident system |
| Messages | Notifications sent to you inside the console. The envelope icon in the top bar opens them. |
| Notification Topic | Which platform events (for example a machine created or deleted, or a sync failure) notify whom |
Configure at least one channel, usually email, before creating alert policies. Otherwise alerts are recorded but reach no one.