Project Phoenix

TRADARS 2.0 Backend Redesign - Complete Architecture Specification

Critical Infrastructure
Central Mainframe Architecture
High Stability Focus

High-Level System Architecture

TIER 1: EXTERNAL DATA SOURCES
Kajabi API
EODHD (Prices)
CFTC (COT Data)
Economic Data APIs
↓
TIER 2: ACQUISITION & VALIDATION LAYER
Kajabi Client
EODHD Client
CFTC Client
FRED/ONS Clients

✓ Rate limiting ✓ Retry logic ✓ Error handling ✓ Credentials secure

↓
TIER 3: DATA INGESTION & TRANSFORMATION

• Data Validation (Integrity checks)

• Schema Transformation (Normalize to Mainframe)

• Deduplication (Upsert by unique key)

• Error Flagging (Invalid data → ErrorLog)

↓
TIER 4: CENTRAL MAINFRAME (Single Source of Truth)

All canonical data:

• Users (from Kajabi)
• Instruments (metadata)
• Market Data (prices)
• Price History (OHLCV)
• Economic Data (indicators)
• COT Data (positioning)
• Analysis Results (radars)
• Cache (pre-computed)
↓
TIER 5: ANALYSIS & PROCESSING LAYER
Day Trading Radar
Swing Trading Radar
Position Trading Radar
Quant Analysis Engine

All read FROM Mainframe, write results BACK to Mainframe

↓
TIER 6: FRONTEND API LAYER (Read-Only)

REST endpoints serving current frontend

Data only from Mainframe (no direct API calls)

Authentication/Authorization enforced

↓
TIER 7: TRADARS 2.0 FRONTEND (Unchanged)
✓ Existing UI remains identical

✓ No changes to user experience

Key Architecture Principles

Single Source of Truth: All data flows through Central Mainframe. No duplication.
Stability First: Robust validation, error handling, and recovery mechanisms.
Observability: All critical operations logged to SystemLog/ErrorLog.
Efficiency: Optimized indexing, caching, and data retrieval.

Critical Issues Being Resolved by Project Phoenix

Kajabi Integration - User Lookup Failures

✓ Resolution: Dedicated Kajabi Client + proper User Mainframe sync ensures all active users found

Risk Radar COT 404 Errors

✓ Resolution: Dedicated CFTC Client with proper error handling, all COT data pre-synced to Mainframe

Zero-Volume IntradayPriceHistory Duplicates

✓ Resolution: Price Validator prevents bad data entry + Data Cleansing Job removes existing bad records

Missing Radar Operation Logging

✓ Resolution: All Radars instrumented with detailed SystemLog entries for every run

QuantAnalysisCache Inconsistency

✓ Resolution: Cache Manager maintains cache integrity after every analysis run

TRADARS v3.0•Published: 9/30/2026 8:10:55 PM