AI-Powered Query Analysis

UptimeDock's AI-Powered Analysis feature uses artificial intelligence to analyze your ClickHouse queries and provide detailed optimization recommendations based on query structure, table schemas, historical performance, and resource usage patterns.

Overview

When investigating slow or resource-intensive queries, understanding the root cause can be challenging. The AI Analysis feature helps by:

• Analyzing your query structure and identifying inefficiencies
• Examining table schemas and indexes
• Reviewing historical performance data
• Providing actionable optimization recommendations

How It Works

When you request an AI analysis, the system:

1. Collects query metadata including execution time, memory usage, and CPU consumption
2. Retrieves relevant table schemas and index information
3. Analyzes historical performance patterns for similar queries
4. Generates a comprehensive analysis with specific recommendations

The analysis is cached, so subsequent views of the same query will display the previously generated analysis instantly.

Accessing AI Analysis

You can access AI Analysis from two locations:

From Query Profile

1. Navigate to your ClickHouse check's Queries tab
2. Click on any query row to open the Query Details modal
3. Select the AI tab
4. Click the Analyze with AI button

From Slow Queries

1. Navigate to your ClickHouse check's Queries tab
2. Switch to the Slow Queries view
3. Click on any slow query to open the Slow Query Details modal
4. Select the AI tab
5. Click the Analyze with AI button

Understanding Analysis Results

The AI analysis provides a structured report with three main sections:

Problem Summary

A concise description of the identified performance issue, including key metrics that indicate the problem. For example:

• Query execution duration
• Number of rows read or written
• Memory allocation
• Affected tables

Root Cause

A detailed explanation of why the query is performing poorly. Common root causes include:

• Missing or ineffective indexes
• Full table scans on large datasets
• Materialized views triggering on every insert
• Suboptimal JOIN operations
• Heavy aggregations without proper filtering
• Default expressions causing additional queries

Recommendations

Actionable optimization suggestions, typically including:

• Query restructuring – How to rewrite the query for better performance
• Index recommendations – Suggested indexes or primary key changes
• Schema changes – Table structure improvements
• Configuration tuning – ClickHouse settings adjustments
• Expected impact – Estimated performance improvement

Best Practices

To get the most value from AI Analysis:

• Analyze high-impact queries first – Focus on queries with high execution counts or longest durations
• Review recommendations in context – Consider your specific use case and constraints
• Test changes in staging – Always validate optimizations before applying to production
• Monitor after changes – Verify improvements using the Query Performance dashboard
• Combine with manual analysis – Use AI insights alongside your own expertise

Important Disclaimer

For complex optimization scenarios or mission-critical systems, we recommend consulting with a database specialist before implementing significant changes.