Guardrails before the bill.

The junior version of cost management is looking at last month's bill and wincing. The engineering version is putting mechanisms in place so the bill cannot surprise you and waste cannot accumulate quietly. Azure gives you three that a cloud engineer is expected to wield: budgets that raise an alert when spend crosses a threshold — before the month closes, not after; Azure Advisor, which continuously inspects your resources and recommends right-sizing and idle-resource cleanup; and storage lifecycle policies, which move cooling data to cheaper tiers and delete it on a schedule without anyone remembering to. This lab sets all three from the command line, so cost control becomes something you deploy, like any other infrastructure.

None of these guardrails costs anything to create — they are governance features, not resources. The only thing you make that could bill is an empty storage account, and empty storage is free.

You cannot optimise a bill after it arrives. You engineer the guardrails before it does.

# Windows/Git Bash: stop it mangling /subscriptions/... arguments (harmless on macOS/Linux)
export MSYS_NO_PATHCONV=1

RG="campux-lab-cost-rg"
az group create -n "$RG" -l eastus

Create a monthly budget scoped to the group, with a notification that emails you when actual spend reaches 80% of the amount. The start date must be the first of a month.

START=$(date -u +%Y-%m-01)

az consumption budget create-with-rg \
  --budget-name campux-monthly \
  --resource-group "$RG" \
  --amount 50 \
  --category Cost \
  --time-grain Monthly \
  --time-period "{\"start-date\":\"$START\",\"end-date\":\"2027-12-31\"}" \
  --notifications "{\"actual-80\":{\"enabled\":true,\"operator\":\"GreaterThanOrEqualTo\",\"threshold\":80.0,\"contact-emails\":[\"[email protected]\"]}}"

Azure Advisor continuously analyses usage and produces cost recommendations — idle resources, oversized VMs, unbought reservations. Read the cost category from the CLI.

az advisor recommendation list --category Cost \
  --query "[].{impact:impact, problem:shortDescription.problem}" -o table

The quiet cost killer is data nobody deletes. A lifecycle policy fixes it structurally: cool blobs after a month, delete them after a year — automatically, forever. Create a storage account and apply one.

SA="campuxcost$RANDOM"
az storage account create -n "$SA" -g "$RG" -l eastus --sku Standard_LRS

cat > policy.json <<'EOF'
{ "rules": [ {
  "name": "tier-and-expire",
  "enabled": true,
  "type": "Lifecycle",
  "definition": {
    "filters": { "blobTypes": ["blockBlob"] },
    "actions": { "baseBlob": {
      "tierToCool": { "daysAfterModificationGreaterThan": 30 },
      "delete":     { "daysAfterModificationGreaterThan": 365 }
    } }
  }
} ] }
EOF

az storage account management-policy create \
  --account-name "$SA" -g "$RG" --policy @policy.json

az storage account management-policy show --account-name "$SA" -g "$RG" \
  --query "policy.rules[0].definition.actions.baseBlob" -o json

Delete the budget explicitly (it is a management object, not a resource in the group), then delete the group.

az consumption budget delete-with-rg --budget-name campux-monthly -g "$RG"
az group delete -n campux-lab-cost-rg --yes
az group exists -n campux-lab-cost-rg      # -> false

You set a Cost Management budget with a threshold alert, queried Azure Advisor's cost recommendations, and applied a storage lifecycle policy that tiers and expires data automatically. That is "drive cost optimization initiatives, including right-sizing and lifecycle policies" — from the job description — implemented as controls rather than described as good intentions. Cost is where cloud engineers earn trust with the people who sign the invoices: a budget that warned the team at 80%, or a lifecycle policy that quietly saved thousands, is the kind of concrete outcome that reads well on a résumé and even better in a review.