Improvement recommendation: AMSP-Win-VMI-CPUUsageOverPercentcritical

Currently, the threshold is hard coded into the query, and the query doesn't look at sustained utilisation over time, so can be very noisy. Recommend the following changes:

Query:
InsightsMetrics
| where Namespace == 'Processor'
and Name == 'UtilizationPercentage'
and Origin == "vm.azm.ms"
| extend Computer=tostring(todynamic(Tags)["vm.azm.ms/computer"])
| summarize AggregatedValue = avg(Val) by bin(TimeGenerated, 15m), _ResourceId

Settings:
Measure = AggregatedValue
AggregationType = Average

Alert Logic
Operator: Greater than
Threshold value = 95