Context Logs & Observability
Grant AI agents real-time visibility into EPM performance, security audits, and integration pipelines.
Why AI Needs Observability
In order for an AI Agent to accurately diagnose a data load failure or explain why a business rule is timing out, it needs access to Oracle's internal logs. Captain-MCP abstracts the complex Interop V1 and V2 APIs into three powerful endpoints designed specifically to feed telemetry data directly into the LLM context window.
Telemetry Endpoints
list_recent_dm_executions
Automatically scans the EPM outbox for recent Data Management process logs. Extracts the Job IDs and file sizes.
↳ Follow up with download_file(logFileName) to read the actual stack trace.
fetch_activity_reports
Discovers and downloads the daily activityreport.json telemetry files. This allows the AI to analyze system performance, average rule duration, and concurrency statistics over time without needing to scrape the UI.
generate_login_reports
Orchestrates Interop V1 to trigger both User Login and Invalid Login CSV reports for a specified date range. Essential for AI-driven security auditing to detect unauthorized access attempts or forgotten passwords.