Skip to main content

Alerting

Alerting turns metrics and events into actionable signals for humans and automation. Unified Monitoring coordinates with your defined policies, routes, channels, maintenance windows, and escalation paths so alerts reach the right people at the right time—without overwhelming them.

This section describes the main building blocks for alerting.

Data Sources for Alerts​

Alerts can be based on:

  • Metrics collected from cloud providers and hosts with Breeze Agent installed.
  • Events generated by Unified Monitoring policies or imported from third‑party tools such as Zabbix, Pingdom, Datadog, and others.
  • Security or compliance findings from other modules.

Design Principles​

When designing alerting:

  • Focus on actionable alerts—every alert should correspond to a clear runbook or escalation path.
  • Prefer symptom‑based alerts (for example, user‑visible errors or saturation) over pure resource utilization when possible.
  • Use CMDB context (application, environment, owner) in routing so alerts reach the teams that can fix them.
  • Use maintenance windows and silences to avoid alert storms during planned work.

For practical guidance, see the playbooks Build Golden Signals and Reduce Alert Noise.

Reading Path​

Policies & Conditions​

Define the rules that evaluate metrics or events and determine when Unified Monitoring should create an alert.

Routes & Channels​

Design routing trees and notification endpoints such as email, Slack, Teams, PagerDuty, or webhooks.

Maintenance & Silences​

Suppress alerts during planned work, known noisy periods, or temporary operational exceptions.

Escalation​

Plan re-notification, handoff, and auto-resolution behavior so alerts continue moving through the right response path.