Microsoft Azure Documentation

Comprehensive guides and resources for Azure services

Monitoring Azure Analysis Services

Overview

Monitoring is crucial for ensuring the health, performance, and availability of your Azure Analysis Services (AAS) instances. Azure provides a comprehensive set of tools and metrics to help you track your service's activity and identify potential issues.

Key aspects to monitor include:

  • Query performance
  • Memory and CPU utilization
  • Refresh operations
  • Connection counts
  • Resource utilization
Azure Analysis Services Monitoring Dashboard Example

Example of a typical Azure monitoring dashboard for AAS.

Key Metrics

Azure Monitor collects a wide range of metrics for Azure Analysis Services. You can view these metrics in the Azure portal, set up alerts based on thresholds, and export them to other services for further analysis.

Commonly Used Metrics:

  • CPU Utilization: Percentage of CPU capacity used by the service.
  • Memory Usage: Amount of memory consumed by the service.
  • Query Count: Number of queries executed against the service.
  • Query Duration: Average duration of executed queries.
  • Data Refresh Duration: Time taken for data refresh operations.
  • Active Connections: Number of active client connections.
  • Successful Data Refreshes: Count of successful data refresh operations.
  • Failed Data Refreshes: Count of failed data refresh operations.

You can access these metrics by navigating to your Analysis Services resource in the Azure portal and selecting "Metrics" from the left-hand menu.

Setting Up Alerts

Alerts notify you when critical conditions are met, allowing you to take proactive measures. You can create alert rules based on specific metrics, event logs, or activity logs.

Creating an Alert Rule:

  1. In the Azure portal, navigate to your Analysis Services resource.
  2. Select "Alerts" from the left-hand menu.
  3. Click on "+ Create" and then "Alert rule".
  4. Configure the Scope to your Analysis Services resource.
  5. In the Condition section, select the Signal name (e.g., "CPU Utilization"). Choose the appropriate Operator (e.g., "Greater than") and set a Threshold value. You can also configure the aggregation granularity and frequency.
  6. In the Actions section, define what happens when the alert is triggered (e.g., send an email, trigger a webhook, run an Automation runbook).
  7. Provide a Details for the alert rule, including a name and severity.
  8. Click "Create alert rule".

For example, you might set up an alert for CPU utilization exceeding 80% for 15 minutes to proactively address performance bottlenecks.

Activity Logs and Diagnostic Settings

Activity logs provide insights into operations performed on your Analysis Services resource, such as creation, deletion, or configuration changes. Diagnostic settings allow you to send these logs, along with metrics, to various destinations for long-term storage and analysis.

Key Log Categories:

  • Administrative: Records management operations like creating or deleting the service.
  • Operational: Captures operational events like query execution and data refreshes.
  • Application: Provides application-specific events and errors.

To configure diagnostic settings:

  1. Navigate to your Analysis Services resource in the Azure portal.
  2. Select "Diagnostic settings" from the left-hand menu.
  3. Click "+ Add diagnostic setting".
  4. Choose the logs and metrics you want to capture.
  5. Select a destination for the logs (e.g., Log Analytics workspace, Storage account, Event Hubs).
  6. Click "Save".

Using Log Analytics with Analysis Services can unlock powerful querying capabilities using Kusto Query Language (KQL).

Troubleshooting Performance Issues

When performance degrades, it's essential to investigate the root cause. The monitoring tools can help identify patterns and anomalies.

  • High CPU/Memory: Investigate complex queries, inefficient model design, or insufficient scaling. Consider optimizing queries, partitioning data, or increasing the service tier.
  • Slow Queries: Analyze query execution plans, identify bottlenecks in the model or DAX, and optimize query patterns.
  • Refresh Failures: Check data source connectivity, permissions, and the size/complexity of the data being refreshed. Review error messages in the activity logs.
  • High Connection Count: Ensure applications are properly managing connections and closing them when no longer needed.

Refer to the Troubleshooting section for more in-depth guidance.