1. The O11 Sanitizer (C# Batch Processing)

Instead of a simple fetch, I developed a custom C# Extension that processes logs in batches.

• Normalisation: Using advanced Regex, the system masks UUIDs, Dates, and Execution Times.
• Fingerprinting: By hashing the normalized message and context, we extract the “DNA” of the error.
• Result: 10,000 ruidous logs collapse into 5 unique logical problems. This is where the real ROI of AI is decided.

/// <summary>
/// (Sneak peak)
/// Normalization by removing patterns to allow log adn grouping
/// </summary>
private string NormalizeString(string input)
{
    if (string.IsNullOrEmpty(input)) return "";

    string result = input;

    result = Regex.Replace(result, @"\d+\s*(ms|secs|seconds|ms)", "[TIME]");

    result = Regex.Replace(result, @"Timer duration = \d+ secs", "Timer duration = [X] secs");
    result = Regex.Replace(result, @"\[retry \d+ of \d+ scheduled\]", "[RETRY]");

    string patternGuid = @"[a-fA-F0-9]{8}-([a-fA-F0-9]{4}-){3}[a-fA-F0-9]{12}";
    result = Regex.Replace(result, patternGuid, "[ID]");

    result = Regex.Replace(result, @"\b\d+\b", "[N]");

    string patternDate = @"\d{2,4}[-/.]\d{2}[-/.]\d{2,4}|\d{2}:\d{2}:\d{2}";
    result = Regex.Replace(result, patternDate, "[DATETIME]");

    return result.Trim();
}

2. The Intelligence Bridge (REST + ODC Agent)

The clean, deduplicated data is sent via REST to an AI Agent running in ODC. This agent isn’t just a generalist; it’s a trained KORE Architect that analyzes the stack trace and provides a structured response:

• Root Cause: Immediate technical origin.
• Fix (O11): Concrete steps to stabilize the current environment.
• ODC Strategy: Strategic warnings (e.g., “This pattern will fail in ODC because of stateless libraries”).

Example: The agent detected a circular dependency pattern in an O11 Service Action and warned that this must be refactored into an ODC Library to maintain the stateless architecture.

3. Human-in-the-Loop Governance

The UI follows a strict State Machine:

`Created` -> `Waiting for Agent` -> `Waiting Acceptance` -> `Validated`.

AI triages the noise and prescribes the solution, but the Architect remains the final governor, validating and accepting the suggestions to ensure quality control.