AZ-305 ยท Logging and monitoring design
Alerting and response design
Metric vs log search vs activity log alerts, action groups and alert processing rules, Service Health alerts, choosing the alert type for a requirement.
Key points
- Metric alerts: precomputed metrics, stateful by default, dynamic thresholds; one rule can cover many resources.
- Log search alerts: a scheduled KQL query.
- Activity log alerts: fire on a control-plane event; stateless.
- Service Health and Resource Health alerts: activity log alerts for platform issues.
- Action groups: reusable receivers (email, SMS, webhook).
- Alert processing rules: add action groups or suppress notifications at scale.
Source: https://learn.microsoft.com/en-us/azure/azure-monitor/alerts/alerts-types
Practice questions
Service Health alerts in Azure Monitor are a kind of which alert type?
- Metric alerts
- Log search alerts
- Activity log alerts
- Smart detection alerts
Show answer
Answer: C. Activity log alerts
Service Health alerts (and Resource Health alerts) are activity log alerts: they fire on Service Health events written to the activity log. They are not based on metrics or on a scheduled KQL query.
Which statements about metric alert rules are true?
- They are stateful by default
- They can use dynamic thresholds
- One rule can cover many VMs in one region
- One rule can cover VMs across several regions
- They can evaluate KQL joins across tables
Show answer
Answer: A. They are stateful by default, B. They can use dynamic thresholds, C. One rule can cover many VMs in one region
Metric alerts are stateful by default, can use machine-learning dynamic thresholds, and one rule can watch many resources of the same type in the same region. A single multi-resource rule cannot span regions, and KQL joins are the territory of log search alerts.
A Service Health alert resolves itself automatically once the Azure incident that triggered it is over.
Show answer
Answer: False
False. Activity log alerts, including Service Health and Resource Health alerts, are event-driven and stateless. The monitor condition stays Fired; you manage the alert by changing its user response to Acknowledged or Closed.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
