Skip to main content

Analyze Service with Manual Entry

For custom applications or services that don't have a pre-built pattern, use the AI Guided Discovery (Manual Analyzer). This workflow allows you to describe your service in plain English, and the system's AI engine (Claude) will help you find, map, and classify it.

Step 1: Define Service Profile​

Start by defining the identity of the service. The Description field is critical—the AI uses it to infer architectural patterns and potential components.

Service Profile Figure 1: Defining the service profile with a descriptive text.

  • Name: Enter a recognizable name (e.g., "Patient Records Portal").
  • Description: Be specific about technologies (e.g., "Java-based web app on Tomcat connecting to Oracle DB").
  • Category: Select the business function (e.g., Healthcare, ERP).

Step 2: Select Known Software​

Search the software catalog to add the components you know are part of this service. This anchors the AI's search.

Software Selection Figure 2: Selecting known software components from the catalog.

Step 3: AI Enrichment & Configuration​

Click Enrich Service. The AI analyzes your description and selected software to:

  1. Suggest Missing Components: It might suggest a web server or cache you missed.
  2. Translate to Processes: It maps software names (e.g., "PostgreSQL") to executable names (e.g., postgres).
  3. Define Tiers: It classifies components into Primary and Dependency tiers.

Enrichment Results Figure 3: Reviewing AI-suggested processes and tier classifications.

Step 4: Visualization & Topology​

Before saving, review the generated topology. The interactive graph shows how the discovered servers communicate.

  • Blue Nodes: Primary application servers.
  • Orange Nodes: Dependent infrastructure (databases, auth servers).

Service Visualization Figure 4: Interactive topology map of the custom service.

Step 5: Review Discovery Results​

The system presents a detailed table of all matched servers. Here you can:

  • Verify Roles: Ensure servers are correctly classified (e.g., web-server vs app-server).
  • Check Confidence: Review the AI's confidence score for each match.
  • Exclude False Positives: Uncheck any servers that don't belong.

Discovery Results Table Figure 5: Validating the list of discovered servers.