Microsoft Azure Architect Design (AZ-301) Practice Exam

What is the appropriate recommendation for generating an alert from Azure Log Analytics when the request charge exceeds a specific threshold?

1. Create a search query to identify when request charge exceeds 50. Configure an alert threshold of 20 and a period of 15.

2. Create a search query to identify when duration exceeds 20 and request charge exceeds 50. Configure a period of 15.

3. Create a search query to identify when request charge exceeds 20. Configure a period of 15 and a frequency of 20.

4. Create a search query to identify when duration exceeds 20. Configure a period of 15.

The correct answer is: Create a search query to identify when request charge exceeds 50. Configure an alert threshold of 20 and a period of 15.

To generate an alert from Azure Log Analytics when the request charge exceeds a specific threshold, the most effective approach is to create a targeted search query that directly correlates with the request charge metric you want to monitor. By crafting a search query that looks for instances where the request charge exceeds a defined limit, and then setting an alert threshold that is appropriately lower than this limit, you can effectively track significant changes in request charges. In this scenario, establishing a threshold of 20 while monitoring for actual spikes at 50 provides a safety margin. This means that even small fluctuations above 20 would trigger awareness of potential performance issues before they escalate. Additionally, configuring a period of 15 ensures that these metrics are evaluated consistently over a reasonable length of time, allowing for better detection of real abnormalities. The choice of monitoring just the request charge—without extraneous metrics—ensures that the alert is directly connected to the specific behavior you want to watch. This focused methodology can help to avoid alert fatigue from notifications that may not be relevant or critical, and it ensures that the alerts are actionable and based on meaningful changes in data. In general, it’s crucial to choose metrics and thresholds that provide a clear signal for intervention without overwhelming the alerting system with noise.