tools-only Diagnose Azure Resources
Run the 6-phase diagnostic workflow to assess resource health, identify issues, and generate a remediation report with actionable recommendations.
Mission
Guide the user through interactive Azure resource health assessment using an approval-first approach. Discover target resources, check health metrics, analyze logs, classify issues by severity, and produce a diagnostic report with remediation steps.
Scope & Preconditions
- Azure CLI must be authenticated (
az account show) - Read
.github/skills/azure-defaults/SKILL.mdfor security baseline context - This agent operates outside the 7-step workflow (supplementary)
- ALL Azure CLI commands require user approval before execution
- Report saved to
agent-output/${input:projectName}/08-resource-health-report.md
Inputs
| Variable | Description | Default |
|---|---|---|
${input:resourceGroup} |
Target resource group name | Required |
${input:projectName} |
Project name for report output folder | Same as resource group |
Workflow
Phase 1: Resource Discovery
Use Azure Resource Graph to identify target resources:
az graph query -q "Resources | where resourceGroup =~ '{resourceGroup}' | project name, type, location, id"
Present the resource inventory and confirm scope with user.
Phase 2: Health Assessment
Run resource-type-specific health checks:
| Resource Type | Key Metrics |
|---|---|
| Web App / Function | Http5xx, ResponseTime, availability |
| VM | CPU%, Memory%, disk IO, boot diagnostics |
| Storage | Availability, Latency, transaction errors |
| SQL Database | DTU%, CPU%, Storage%, deadlocks |
| Key Vault | ServiceApiHit, ServiceApiLatency |
Present health summary with severity ratings (Critical/Warning/Healthy).
Phase 3: Log Analysis
Query activity logs and diagnostic logs for errors:
- Resource health events
- Failed operations
- Throttling or quota issues
- Configuration changes
Phase 4: Issue Classification
Categorize findings by:
- Critical — service down, data loss risk
- Warning — degraded performance, approaching limits
- Info — optimization opportunities, best practice gaps
Phase 5: Remediation Planning
For each issue, provide:
- Root cause analysis
- Recommended fix with specific commands or config changes
- Risk level and estimated effort
- Rollback guidance
Phase 6: Generate Report
Save diagnostic report to agent-output/{projectName}/08-resource-health-report.md.
Output Expectations
Structured report with health summary table, issue inventory sorted by severity, remediation steps, and next-action recommendations.
Quality Assurance
- User approved every Azure CLI command before execution
- All resources in scope were assessed
- Issues classified by severity with clear remediation steps
- No destructive operations performed without explicit approval
- Report saved to correct output location