Open YouTube

Introduction to Azure Monitor | AZ-104 | Episode 27

autoenpublicupdated Jul 5, 2026, 18:06

Why monitoring matters

Building and recovering a solution are not enough; operations also require confirming that the solution meets expected business outcomes and performance requirements.
Azure Monitor provides tools to measure expected outcomes, tune performance, and support business decisions with data.
Metrics need context and baselines because raw numbers alone do not explain whether a condition is good, bad, normal, or abnormal.

Azure Monitor acts as a central service for collecting telemetry across Azure resources, on-premises systems, and other cloud services.
Microsoft already monitors Azure infrastructure for health, recovery, and billing, and Azure Monitor exposes telemetry so customers can make their own operational decisions.
Telemetry can come from tenant-level data, subscriptions, Azure resources, operating systems, containers, infrastructure workloads, and custom applications.

The Azure Monitoring Agent can extend monitoring to customer-controlled environments such as on-premises devices.
Tenant and subscription data are available by default, but multiple tenants may require aggregation tools such as Azure Lighthouse or Log Analytics Workbooks.
Azure can see resource-level signals such as VM CPU usage, but detailed workload activity inside a VM, container, or application must be emitted by the customer’s configuration.
Application Insights enables custom application telemetry such as page visits, regions, IP addresses, button clicks, and page load times.

Azure Monitor data primarily appears as metrics and logs.
Metrics are numerical values emitted near real time, but they must be evaluated against baselines and related context.
Logs are event-triggered records that collect a set of diagnostic details at a point in time.
Traces show request paths through a system, adding another view beyond isolated events or metric values.

Azure Monitor supports four major response areas: insights, visualization, analysis, and response.
Insights help identify performance patterns and whether scaling or other changes may be needed.
Visualization can be delivered through dashboards, Power BI, Grafana, webpages, PDFs, or emails for different audiences.
Log Analytics and related tools help query and analyze collected data to identify trends or problems.
Alerts and automation can trigger responses ranging from emails to complex workflows such as service ticket creation and vendor coordination.

Azure logs can be queried with Kusto Query Language, but the portal also provides filters for users who do not know KQL.
Azure stores log information for 90 days by default for free.
Because breaches may be discovered long after 90 days, organizations may need longer retention in a Log Analytics workspace.
Retention length requires judgment because keeping too little data can limit investigation, while keeping too much can increase cost and make analysis harder.

The Azure portal shows Activity Logs and Azure Monitor features on resource blades, including logs, metrics, alerts, diagnostic settings, and workbooks.
Activity Logs describe actions on a resource, such as backup configuration or restore point creation, not necessarily activity inside the resource itself.
Event Viewer on a local Windows machine illustrates how logs capture detailed information for events that occurred at specific points in time.
Performance counters illustrate metrics, but collecting every counter can create excessive and confusing data.
Administrators should choose the correct metrics, validate them, set baselines, and measure deviations from those baselines.

Azure Monitor centralizes metrics, logs, traces, and related telemetry so organizations can compare systems, gain operational insight, and respond to maintain business results.