feat: Add comprehensive logs management tools

[aficiomaquinas]

📋 Add Comprehensive Logs Management Tools

🎯 Overview

This PR adds a complete set of logs management tools to the Kestra MCP server, enabling users to easily access, search, and manage execution logs directly through the MCP interface.

✨ What's New

6 New Logs Tools Added:

• get_execution_logs - Retrieve logs for a specific execution with filtering options
• download_execution_logs - Download logs as plain text for easy analysis
• search_logs - Global log search across all executions with advanced filters
• delete_execution_logs - Delete logs for a specific execution (use with caution)
• delete_flow_logs - Delete logs for all executions of a specific flow
• follow_execution_logs - Real-time log streaming for running executions

🔍 Advanced Filtering Capabilities:

• Log levels: ERROR, WARN, INFO, DEBUG, TRACE
• Task filtering: taskId, taskRunId, attempt number
• Date ranges: For search operations (startDate, endDate)
• Namespace and flow filtering: Targeted log searches

🚀 Key Benefits

• Easy Log Access: No need to navigate to Kestra UI for log inspection
• Powerful Search: Find specific logs across all executions quickly
• Real-time Monitoring: Follow logs as they're generated during execution
• Flexible Export: Download logs in various formats for analysis
• Complete Coverage: Works with both Community and Enterprise editions

📚 Documentation & Testing

• ✅ Complete documentation in docs/prompts/logs.md
• ✅ Comprehensive test suite with 6 test cases
• ✅ Integration tests with real Kestra instance
• ✅ Parameter validation and error handling
• ✅ All tests passing (33/33 tests pass)

🔧 Technical Implementation

• Uses official Kestra REST API endpoints for logs
• Compatible with both Basic Auth and API Token authentication
• Proper error handling and input validation
• Follows existing MCP server patterns and conventions
• No breaking changes to existing functionality

📊 Impact

This contribution significantly improves the developer experience by providing:

• Direct log access through MCP interface
• Reduced context switching between tools
• Enhanced debugging capabilities for Kestra workflows
• Better log management for production environments

🧪 Testing

All tests pass successfully:

$ pytest tests/test_logs.py -v
======================== 6 passed in 21.68s =========================

📝 Usage Examples

# Get logs for a specific execution
Get logs for execution abc123 with INFO level or higher

# Search for errors across executions
Search for error logs in namespace production.etl

# Download logs for analysis
Download all logs from execution xyz789

# Follow logs in real-time
Follow ERROR level logs for execution abc123

🔗 API Endpoints Used

• GET /api/v1/{tenant}/logs/{executionId} - Get execution logs
• GET /api/v1/{tenant}/logs/{executionId}/download - Download logs
• GET /api/v1/{tenant}/logs/search - Search logs globally
• DELETE /api/v1/{tenant}/logs/{executionId} - Delete execution logs
• DELETE /api/v1/{tenant}/logs/{namespace}/{flowId} - Delete flow logs
• GET /api/v1/{tenant}/logs/{executionId}/follow - Follow logs (SSE)

📁 Files Changed

• src/tools/logs.py - Main implementation (335 lines)
• docs/prompts/logs.md - Complete documentation (88 lines)
• tests/test_logs.py - Test suite (208 lines)
• src/server.py - Tool registration (3 lines)
• tests/test_mcp_server.py - Updated tool list (7 lines)
• README.md - Added logs to available tools (1 line)

Total: 6 files changed, 642 insertions(+)

---

This PR makes log management much more accessible and powerful for Kestra users, providing a seamless experience for debugging and monitoring workflows through the MCP interface.

[MilosPaunovic]

@loicmathieu I'm not sure if you're the right person for a review here, if not, please re-assign.