Building Azure Monitoring Dashboards with MCP and Terraform

Overview

This guide demonstrates how to create comprehensive Azure monitoring dashboards using Infrastructure as Code (IaC) with Terraform, enhanced by the Model Context Protocol (MCP) for Azure provider documentation. We'll walk through building a complete Container App monitoring solution with dashboards, alerts, and log analytics integration.

Prerequisites
Architecture Overview
Setting Up MCP for Azure
Terraform Dashboard Configuration
Dashboard Template Design
Monitoring Components
Deployment Process
Best Practices
Troubleshooting

Prerequisites

Azure subscription with appropriate permissions
Terraform >= 1.4.0
Azure CLI configured
VS Code with MCP support
Understanding of Azure Container Apps, Log Analytics, and Azure Monitor

Architecture Overview

Our monitoring infrastructure consists of:

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Container App │────│  Log Analytics   │────│ Azure Monitor   │
│                 │    │   Workspace      │    │    Alerts       │
└─────────────────┘    └──────────────────┘    └─────────────────┘
         │                        │                       │
         └────────────────────────┼───────────────────────┘
                                  │
                         ┌─────────────────┐
                         │ Azure Portal    │
                         │   Dashboard     │
                         └─────────────────┘

Setting Up MCP for Azure

1. Configure Azure Provider Documentation Access

The Model Context Protocol (MCP) provides real-time access to Azure provider documentation, ensuring we use the latest syntax and best practices.

// Example MCP configuration for Azure Terraform provider
const azureProviderMCP = {
  provider: "hashicorp/azurerm",
  version: "~> 4.0",
  namespace: "hashicorp",
  capabilities: ["resources", "data-sources", "guides"]
};

2. Key MCP Benefits

Real-time Documentation: Access latest Azure provider syntax
Resource Discovery: Find available metrics and properties
Version Compatibility: Ensure correct provider versions
Best Practices: Get recommendations for resource configuration

Terraform Dashboard Configuration

1. Provider Setup

terraform {
  required_providers {
    azurerm = {
      source  = "hashicorp/azurerm"
      version = "~> 4.0"
    }
  }
}

provider "azurerm" {
  features {}
}

2. Core Dashboard Resource

resource "azurerm_portal_dashboard" "container_app_monitoring" {
  name                = local.dashboard_name
  resource_group_name = local.resource_group_name
  location            = data.terraform_remote_state.base.outputs.location
  
  tags = merge(local.tags, {
    "hidden-title" = "Container App Monitoring - ${title(var.ENVIRONMENT)}"
    "monitoring"   = "enabled"
  })

  dashboard_properties = templatefile("${path.module}/dashboard.tpl", {
    container_app_id             = azurerm_container_app.main.id
    container_app_name           = azurerm_container_app.main.name
    log_analytics_workspace_id   = local.log_analytics_workspace_id
  })
}

3. Variables Configuration

variable "enable_monitoring_alerts" {
  description = "Enable monitoring alerts for the container app"
  type        = bool
  default     = true
}

variable "cpu_threshold_nanocores" {
  description = "CPU threshold in nanocores for alerting"
  type        = number
  default     = 500000000  # 50% of 1 vCPU
}

variable "memory_threshold_bytes" {
  description = "Memory threshold in bytes for alerting"
  type        = number
  default     = 536870912  # 512MB
}

Dashboard Template Design

1. Template Structure

Create a separate dashboard.tpl file for maintainability:

{
  "lenses": {
    "0": {
      "order": 0,
      "parts": {
        "0": {
          "position": {
            "x": 0, "y": 0,
            "rowSpan": 4, "colSpan": 12
          },
          "metadata": {
            "type": "Extension/HubsExtension/PartType/MarkdownPart",
            "settings": {
              "content": {
                "settings": {
                  "content": "# 📊 Container App Monitoring Dashboard\n\n**Container App**: ${container_app_name}",
                  "title": "Container App Monitoring"
                }
              }
            }
          }
        }
      }
    }
  }
}

2. Layout Strategy

Our dashboard uses a structured layout:

Header Panel (rowSpan: 4) - Overview and navigation links
Error Metrics (full width) - Priority monitoring
Performance Metrics (2x2 grid) - CPU, Memory, Requests, Replicas

3. Navigation Integration

**Monitoring Features:**
- ❌ **Error Metrics** → 🔍 **[View Error Logs](https://portal.azure.com/#@${log_analytics_workspace_id}/blade/Microsoft_Azure_Monitoring_Logs/LogsBlade)**
- 📈 **CPU Usage** - Real-time processor performance metrics
- 💾 **Memory Usage** - RAM consumption monitoring
- 🌐 **Request Rate** - Traffic patterns
- 🔄 **Replica Count** - Auto-scaling behavior
- ⚠️ **Automated Alerts** → 🚨 **[View Alerts](https://portal.azure.com/#blade/Microsoft_Azure_Monitoring/AzureMonitoringBrowseBlade/~/alerts)**

Monitoring Components

1. Metric Alerts

resource "azurerm_monitor_metric_alert" "cpu_usage" {
  count               = var.enable_monitoring_alerts ? 1 : 0
  name                = "${azurerm_container_app.main.name}-cpu-usage-alert"
  resource_group_name = local.resource_group_name
  scopes              = [azurerm_container_app.main.id]
  description         = "Alert when CPU usage is high"
  severity            = 2
  frequency           = "PT5M"
  window_size         = "PT5M"
  enabled             = true
  auto_mitigate       = true

  criteria {
    metric_namespace = "Microsoft.App/containerApps"
    metric_name      = "UsageNanoCores"
    aggregation      = "Average"
    operator         = "GreaterThan"
    threshold        = var.cpu_threshold_nanocores
  }
}

2. Log-Based Alerts

resource "azurerm_monitor_scheduled_query_rules_alert_v2" "application_errors" {
  count               = var.enable_monitoring_alerts ? 1 : 0
  name                = "${azurerm_container_app.main.name}-application-errors"
  resource_group_name = local.resource_group_name
  location            = data.terraform_remote_state.base.outputs.location

  evaluation_frequency = "PT5M"
  window_duration      = "PT10M"
  scopes               = [local.log_analytics_workspace_id]
  severity             = 1
  enabled              = true
  auto_mitigation_enabled = true

  criteria {
    query = <<-QUERY
      ContainerAppConsoleLogs_CL
      | where ContainerAppName_s == "${azurerm_container_app.main.name}"
      | where TimeGenerated > ago(10m)
      | where Log_s contains "ERROR" or Log_s contains "Exception"
      | summarize count()
    QUERY
    time_aggregation_method = "Count"
    threshold               = var.error_log_threshold
    operator                = "GreaterThan"
  }
}

3. Diagnostic Settings

resource "azurerm_monitor_diagnostic_setting" "container_app" {
  name                       = "${azurerm_container_app.main.name}-diagnostics"
  target_resource_id         = azurerm_container_app.main.id
  log_analytics_workspace_id = local.log_analytics_workspace_id

  enabled_metric {
    category = "AllMetrics"
  }
}

Deployment Process

1. Environment Configuration

Create environment-specific configuration in env.development.json:

{
  "enable_monitoring_alerts": true,
  "cpu_threshold_nanocores": 500000000,
  "memory_threshold_bytes": 536870912,
  "request_rate_threshold": 5000,
  "error_log_threshold": 1
}

2. Deployment Commands

# Initialize Terraform
terraform init

# Plan the deployment
terraform plan -var-file=env.development.json

# Apply the changes
terraform apply -var-file=env.development.json

3. Verification

After deployment, verify:

Dashboard appears in Azure Portal
All panels display data correctly
Navigation links work properly
Alerts are configured and active

Best Practices

1. Template Organization

Separate Templates: Use external .tpl files for complex dashboard JSON
Parameterization: Use variables for all configurable values
Validation: Include template validation in CI/CD pipeline

2. Monitoring Configuration

Environment-Specific Thresholds: Different alerts for dev/staging/production
Auto-Mitigation: Enable automatic alert resolution
Comprehensive Coverage: Monitor both infrastructure and application metrics

3. Dashboard Design

Logical Layout: Group related metrics together
Priority Placement: Put critical metrics (errors) at the top
Navigation: Include direct links to relevant Azure services
Documentation: Provide KQL queries for manual investigation

4. Infrastructure as Code

Version Control: All dashboard configurations in Git
Modular Design: Reusable modules for different applications
State Management: Proper Terraform state management
Documentation: Comprehensive README and inline comments

Troubleshooting

Common Issues

Dashboard Not Appearing

# Check resource group and tags
az portal dashboard list --resource-group <rg-name>

Links Not Working
- Verify URL format: #@${workspace_id}/blade/...
- Check workspace ID format and permissions
Metrics Not Displaying
- Verify diagnostic settings are enabled
- Check Container App is generating metrics
- Ensure Log Analytics workspace connection
Provider Version Issues
- Use MCP to verify latest Azure provider syntax
- Check for breaking changes in provider updates
- Test with specific provider versions

Debugging Commands

# Check Terraform plan
terraform plan -detailed-exitcode

# Validate template
terraform validate

# Check Azure resources
az container app show --name <app-name> --resource-group <rg-name>
az monitor log-analytics workspace show --resource-group <rg-name> --workspace-name <workspace-name>

Conclusion

This approach combines the power of Infrastructure as Code with real-time Azure documentation access through MCP. Key benefits include:

Consistency: Standardized monitoring across environments
Maintainability: Version-controlled dashboard configurations
Scalability: Reusable templates for multiple applications
Best Practices: MCP ensures latest Azure provider features
Automation: Integrated deployment pipeline

The result is a professional, maintainable monitoring solution that provides comprehensive observability for Azure Container Apps with minimal manual configuration.

Additional Resources

This guide demonstrates a production-ready Container App monitoring solution, showcasing enterprise-level monitoring with Infrastructure as Code best practices.

Lorenzo GM

Building Azure Monitoring Dashboards with MCP and Terraform

Building Azure Monitoring Dashboards with MCP and Terraform

Overview

Table of Contents

Prerequisites

Architecture Overview

Setting Up MCP for Azure

1. Configure Azure Provider Documentation Access

2. Key MCP Benefits

Terraform Dashboard Configuration

1. Provider Setup

2. Core Dashboard Resource

3. Variables Configuration

Dashboard Template Design

1. Template Structure

2. Layout Strategy

3. Navigation Integration

Monitoring Components

1. Metric Alerts

2. Log-Based Alerts

3. Diagnostic Settings

Deployment Process

1. Environment Configuration

2. Deployment Commands

3. Verification

Best Practices

1. Template Organization

2. Monitoring Configuration

3. Dashboard Design

4. Infrastructure as Code

Troubleshooting

Common Issues

Debugging Commands

Conclusion

Additional Resources