Exolar Integration

Fetching CI analytics and test history from Exolar for intelligent failure diagnosis.

Overview

Exolar is a CI analytics database that stores test results from your CI/CD pipeline. The AI uses Exolar to:

• Fetch historical test results and trends
• Query failure patterns and clusters
• Analyze flake rates and test reliability
• Compare current failures against past occurrences

Important: Data flows FROM your CI pipeline TO Exolar. The AI READS from Exolar to make informed decisions - it does not report to Exolar.

Data Flow

┌─────────────┐         ┌─────────────┐         ┌─────────────┐
│   CI/CD     │────────▶│   EXOLAR    │◀────────│   Claude    │
│  Pipeline   │  Pushes │  Database   │  Queries│   (AI)      │
│             │  Results│             │  Data   │             │
└─────────────┘         └─────────────┘         └─────────────┘

Authentication

On first use, authenticate at:

https://exolar.ai-innovation.site/

Available Tools

query_exolar_data

Query test data and analytics. This is the primary tool for fetching information.

Find similar failures:

query_exolar_data({
  dataset: "failures",
  filters: {
    error_pattern: "timeout waiting for selector",
    project: "my-project"
  }
})

Check flake history:

query_exolar_data({
  dataset: "flaky_tests",
  filters: {
    test_file: "login.spec.ts",
    time_range: "30d"
  }
})

Get test trends:

query_exolar_data({
  dataset: "test_results",
  filters: {
    project: "my-project",
    time_range: "7d"
  }
})

Find tests by classification:

query_exolar_data({
  dataset: "classified_failures",
  filters: {
    classification: "FLAKE",
    resolved: false
  }
})

Common Query Patterns

Failure Classifications

When analyzing failures, Exolar uses these classifications:

Workflow Integration

When Tests Fail

1. Fetch similar failures:

   query_exolar_data({
     dataset: "failures",
     filters: { error_pattern: "Element not found" }
   })
   

2. Check if known flake:

   query_exolar_data({
     dataset: "flaky_tests",
     filters: { test_name: "login should work" }
   })
   

3. Verify with Playwright MCP:

   • Navigate to the app
   • Inspect actual state
   • Compare expected vs actual

4. Classify and act:

   • Use triqual-plugin:failure-classifier agent
   • Apply appropriate fix based on classification

Learning from History

The triqual-plugin:pattern-learner agent uses Exolar data to:

1. Identify recurring failure patterns
2. Propose documentation updates to Quoth
3. Improve future test generation

Analytics Dashboard

View your test analytics at:

https://exolar.ai-innovation.site/dashboard

Features:

• Test pass/fail trends
• Flake rate tracking
• Failure clustering by error type
• Test duration analysis
• Historical comparison

Best Practices

1. Always query before diagnosing - Check if failure is known
2. Use filters effectively - Narrow down to relevant data
3. Combine with Playwright MCP - Verify app state, don't assume
4. Trust classifications - Past classifications inform current analysis
5. Look for patterns - Recurring errors suggest systematic issues