AZ-104 domain - 14% of the exam

Monitor and Maintain Azure Resources

Monitor and Maintain Azure Resources is 14% of the AZ-104 exam. These are the objectives it covers, each with practice questions, with every answer explained.

The domain in numbers

34
Practice questions
2
Objectives
28
Medium
6
Hard

Where people go wrong

  • Giving a tool the wrong kind of data

    The wrong answers assign a data type to the wrong Azure Monitor tool. Metrics Explorer charts numeric metrics but does not run KQL over log tables. Logs holds log and trace records queried with KQL, not a time-series numeric store. The activity log records control-plane operations, and Service Health reports platform incidents, so neither stores performance values. Metrics suit numeric values at regular intervals, Log Analytics in KQL mode suits read-only queries, and an activity log alert suits a discrete event such as a role assignment.

    See it in a sample question
  • Crediting a diagnostic setting with abilities it lacks

    Storage and a Log Analytics workspace are offered for streaming to a non-Microsoft SIEM, although only Event Hubs forwards a stream to external systems. One setting is credited with two workspaces, although a setting defines at most one of each destination type, so a second workspace needs a second setting. The wrong options also cap a resource at three settings when the limit is five, demand the same subscription when cross-subscription destinations are allowed, and demand premium storage, which is not supported.

    See it in a sample question
  • Swapping Next hop and IP flow verify

    The wrong answers hand routing questions to IP flow verify and NSG diagnostics, and filtering questions to Next hop. IP flow verify checks whether a packet is allowed or denied at the virtual machine and names the security rule responsible. Next hop reports routing: the hop type, IP address and route table ID for a destination. Connection monitor, Packet capture and Topology are also offered, yet none of them names the deciding rule or computes a route.

    See it in a sample question
  • Treating alert rules, processing rules and action groups as one layer

    An alert rule combines the monitored resources, the signal and the conditions; action groups and suppression sit downstream of it. Alert processing rules do not create alerts, they only modify alerts that are already firing, and their scope stays within the same subscription. When a suppression rule and an apply-action-groups rule match one alert, suppression takes precedence. The wrong options claim the two cancel out, or that applying action groups wins.

    See it in a sample question
  • Blurring drills, failover and failback in Site Recovery

    Disaster recovery drills validate a plan without affecting ongoing replication. Planned failover, unplanned failover and failback each move or return a workload, so offering one as a non-disruptive test fails. Planned failover is for expected outages with zero-data loss; unplanned failover answers a disaster with minimal loss that depends on replication frequency. The wrong options also claim failback is blocked after failover, or that a planned failover needs a fully lost region.

    See it in a sample question

Objectives in this domain

Sample questions from this domain

Free sampleMonitor and Maintain Azure Resourcesmedium

An administrator wants to be notified whenever a new role assignment is written at the subscription scope. The trigger is an event recorded in the subscription log of control-plane operations, not a numeric measurement. Which alert type should the administrator configure?

  • AMetric alert that samples a resource counter at set intervals
  • BActivity log alert that fires when a matching log event occurs Correct
  • CSmart detection alert that flags anomalies in a web application
  • DPrometheus alert that evaluates a stored PromQL query result
Use an activity log alert when the trigger is a discrete subscription event, and a metric alert when the trigger is a sampled numeric value. Activity log alerts are triggered when a new activity log event occurs that matches the defined conditions, which is exactly how a control-plane operation such as a role assignment is captured.

Why A is wrong: Metric alerts evaluate resource metrics at regular intervals, so they suit numeric measurements rather than a discrete control-plane event.

Why B is correct: Correct. Activity log alerts are triggered when a new activity log event occurs that matches the defined conditions, which is exactly how a control-plane operation such as a role assignment is captured.

Why C is wrong: Smart detection warns of performance problems and failure anomalies in an Application Insights web application, not subscription role changes.

Why D is wrong: Prometheus alerts target Prometheus metrics using PromQL, which does not match a discrete activity log event.

Free sampleMonitor and Maintain Azure Resourcesmedium

A security team must stream resource logs from an Azure resource to a non-Microsoft SIEM platform in near real time. Which diagnostic setting destination should you choose for this streaming requirement?

  • AAn Azure Event Hubs namespace that forwards the stream to external systems Correct
  • BAn Azure Storage account that retains the records for audit and backup
  • CA Log Analytics workspace that holds the records for KQL log queries
  • DAn Azure Monitor workspace that holds Prometheus metrics for PromQL
Event Hubs is the diagnostic settings destination for streaming logs to external or non-Microsoft systems. Event Hubs is the diagnostic setting destination for streaming data to external systems such as non-Microsoft SIEM solutions, which is exactly the streaming integration the team needs.

Why A is correct: Correct. Event Hubs is the diagnostic setting destination for streaming data to external systems such as non-Microsoft SIEM solutions, which is exactly the streaming integration the team needs.

Why B is wrong: Storage is for retention and static analysis, not near real-time streaming to a SIEM.

Why C is wrong: A workspace supports KQL analysis and log alerts, not streaming out to an external SIEM.

Why D is wrong: An Azure Monitor workspace stores Prometheus metrics and is not a diagnostic settings log destination.

Free sampleMonitor and Maintain Azure Resourcesmedium

A virtual machine cannot reach an external IP address and the administrator suspects a security rule is blocking the packet. They want a tool that confirms whether a packet is allowed or denied at the VM and names the rule responsible. Which Network Watcher tool should they use?

  • ANext hop checks the route and names the rule
  • BConnection monitor checks the packet and names the rule
  • CIP flow verify checks the packet and names the rule Correct
  • DPacket capture checks the packet and names the rule
IP flow verify answers allowed-or-denied at the VM and names the deciding security rule, while next hop answers routing questions. IP flow verify detects traffic filtering issues at the virtual machine level, checks whether a packet is allowed or denied to or from an IP address, and tells you which security rule made that decision.

Why A is wrong: Next hop detects routing issues and returns the hop type and route table, but it does not evaluate whether a security rule allows or denies a packet.

Why B is wrong: Connection monitor provides end-to-end connectivity monitoring over time and does not identify the specific security rule blocking a packet.

Why C is correct: Correct. IP flow verify detects traffic filtering issues at the virtual machine level, checks whether a packet is allowed or denied to or from an IP address, and tells you which security rule made that decision.

Why D is wrong: Packet capture records traffic to and from a VM for later analysis but does not state which security rule allowed or denied a packet.

Free sampleMonitor and Maintain Azure Resourcesmedium

An engineer is reviewing how an alert rule decides to fire. The rule watches data from a resource, captures a signal, and compares it against criteria. Which combination does an alert rule bring together to make that decision?

  • AThe action group, the schedule, and the suppression scope
  • BThe monitored resources, the signal, and the conditions Correct
  • CThe severity, the user response, and the retention period
  • DThe dynamic threshold, the workspace, and the dimension
An alert rule is defined by resource plus signal plus condition, and it triggers the action group only when the condition is met. An alert rule combines the resources to be monitored, the signal or data from the resource, and the conditions, then fires an alert when the conditions are met.

Why A is wrong: Action groups and suppression are downstream of the rule, and scheduling belongs to alert processing rules rather than the alert rule itself.

Why B is correct: Correct. An alert rule combines the resources to be monitored, the signal or data from the resource, and the conditions, then fires an alert when the conditions are met.

Why C is wrong: Severity, user response, and retention describe alert state and lifecycle, not the inputs the rule combines to evaluate a signal.

Why D is wrong: These are configuration details of specific alert types rather than the three-part definition every alert rule combines.

Free sampleMonitor and Maintain Azure Resourcesmedium

Your team wants to validate the Site Recovery disaster recovery plan for a multi-tier app without affecting the ongoing replication of the production workload. Which Site Recovery capability supports this?

  • APlanned failover that switches production over to the secondary site
  • BUnplanned failover that activates the secondary site after a disaster
  • CFailback that returns the workload to the primary site after recovery
  • DDisaster recovery drills that run without affecting ongoing replication Correct
Site Recovery drills validate the recovery plan without disrupting ongoing replication; failover and failback move the live workload. Site Recovery supports testing without disruption, letting you easily run disaster recovery drills without affecting ongoing replication of the protected workload.

Why A is wrong: A planned failover actually moves the workload to the secondary location for an expected outage, rather than validating without disruption.

Why B is wrong: An unplanned failover responds to an unexpected disaster with minimal data loss; it is not a non-disruptive validation of the plan.

Why C is wrong: Failback returns the workload to the primary site once it is available again, which is not a test of the plan.

Why D is correct: Correct. Site Recovery supports testing without disruption, letting you easily run disaster recovery drills without affecting ongoing replication of the protected workload.

Other domains in this exam

See also the AZ-104 cert hub, the study guide, and the cheat sheet.

Examworthy is not affiliated with or endorsed by Microsoft. Original, blueprint-aligned practice material only.