Learn

📌 Problem You need to monitor, analyze, and control the usage and performance of Azure AI services to optimize cost, detect issues, and ensure service health.

✅ Solution with Azure Use Azure Monitor, Cost Management, Alert Rules, Diagnostic Logging, and Dashboards to monitor usage, manage costs, and diagnose problems in Azure AI services.

🧩 Required Components * Azure Pricing Calculator * Azure Cost Management + Billing * Azure Monitor * Metrics and Alerts * Azure Log Analytics workspace * Azure Storage account * Diagnostic settings configuration * Azure Dashboards

🏗️ Architecture / Development

🧾 Monitor costs * Use Azure Pricing Calculator to estimate costs before deployment. * Select service (e.g., Azure AI Language), region, tier, usage metrics. * Save/export estimate in Excel. * Use Cost Analysis in Azure portal to monitor actual costs. * Apply filters to see only Azure AI Services usage.

🚨 Create alerts * Navigate to your resource → Alerts tab → Add alert rule. * Define: * Scope: AI service resource * Condition: Metric (e.g., error count) or Activity Log (e.g., key regeneration) * Action: Email, Logic App, etc. * Alert details: Name, resource group

📊 View metrics * Use Metrics blade on AI resource to: * Add custom charts (e.g., call count, error rate) * Export or link charts * Create Dashboards with: * Up to 100 tiles combining multiple visualizations * Add metric charts directly from resources

🧾 Manage diagnostic logging * Create storage destinations: * Azure Log Analytics (for querying/log visualization) * Azure Storage (for archival/export) * Configure on Diagnostic settings: * Define name, categories of logs/metrics, destinations * View diagnostic data: * Wait up to 1 hour for data flow * Use Log Analytics query interface to analyze logs

🧠 Best Practices / Considerations * Use free tier for testing to reduce costs. * Always create estimates before provisioning. * Set up alerts for error thresholds and key operations. * Use dashboards for centralized monitoring. * Configure diagnostics to send logs to both Log Analytics and Storage. * Ensure storage accounts are in the same region as AI resources for latency/cost efficiency.

❓ Simulated Exam Questions 1. 🧮 How can you estimate costs for multiple Azure AI services before deployment? ➤ Use the Azure Pricing Calculator, select services and input expected usage. 2. 💰 Where can you view the accumulated cost of Azure AI services in the portal? ➤ Use Cost Analysis and filter by service name = "Azure AI Services". 3. 🚨 What types of signals can trigger an alert rule in Azure AI services? ➤ Activity Log (e.g., key regeneration), Metric (e.g., error count > 10). 4. 📊 How do you add service metrics to a shared view? ➤ Use Metrics → add chart → Add to Dashboard. 5. 🛠️ Which resources must be created before enabling diagnostic logging? ➤ Azure Log Analytics and/or Azure Storage. 6. 🔍 Where can you analyze captured diagnostic data for an AI resource? ➤ In Azure Log Analytics, using query interface.