Candidates who only read or watch tend to struggle with AZ-104 scenario questions, because those questions describe a situation and ask what you would do. People who have configured, broken and repaired the service answer them quickly.
A weekly rhythm that works
- Plan (30 minutes). Pick the skill area and write down what you expect to build.
- Build (90 minutes). Do it in the portal first, then repeat it with the CLI or a template.
- Break (30 minutes). Remove a route, tighten a firewall rule, delete a role assignment — then diagnose it.
- Explain (15 minutes). Write two sentences on why it works, as if teaching someone else.
Labs worth repeating
- A hub-and-spoke network with peering and a network security group.
- A storage account locked down with a firewall, a private endpoint and a SAS token.
- A VM scale set behind a load balancer with an autoscale rule.
- A management-group hierarchy with a policy that enforces tagging.
- Azure Backup for a VM, followed by a restore test.
Watch the costs
Set a budget and alert on your subscription, and delete resource groups when you finish. Deallocating VMs stops compute charges, but disks and public IPs continue to cost money.
Validate everything
The habit that separates strong administrators is validation: after every change, prove it works with a test — a ping, a connection test in Network Watcher, a failed sign-in you expected. That habit is also what our labs reward.