Lending Journey

📌 End-to-End Parallel Lending Journey (Text Architecture Flow)

1. Login / Authentication

• Persona: Customer (Parag Shah)

• Action: Logs in to Lending App

• Flow:

  1. Frontend (Web/Mobile) → Azure API Management → Auth Microservice (AKS).

  2. Auth MS validates Parag’s credentials with Azure AD B2C.

  3. JWT token issued, user profile fetched from PostgreSQL BDR (Customer table).

  4. Redis Cache Write:

     • Cache user profile: customer:{parag_id} → {name, email, mobile, KYC status, loan summary}

  5. Kafka Topic:

     • Publish CustomerLoginEvent → consumed by Customer360 MS to update last login timestamp in PostgreSQL.

2. Loan Application Initiation

• Persona: Parag Shah

• Action: Starts loan application (selects product, amount)

• Flow:

  1. Loan Application MS writes to PostgreSQL BDR → loan_applications table (status=“Initiated”).

  2. Publishes LoanInitiatedEvent to Kafka.

  3. Consumers:

     • KYC MS → triggers KYC process.

     • Customer360 MS → updates loan summary in Cosmos DB.

  4. Cosmos DB Insert:

     { "eventId": "evt-101", "customerId": "cust-001", "loanAppId": "loan-101", "eventType": "LoanInitiated", "timestamp": "2025-08-30T10:01:00Z", "payload": { "amount": 500000, "tenure": 36 } }

  5. Redis Cache Update:

     • Update cache key customer:{parag_id}:loan_summary with {loanAppId, status=Initiated}.

Step 4 – Parallel Execution: KYC + Credit Score + Fraud Check

• Action: After consent + document upload, orchestrator publishes events in parallel.

• Kafka Topics Fired Simultaneously:

  • kyc-check-requested

  • credit-score-requested

  • fraud-check-requested

Consumers:

Topic	Consumer Service	Action	Cache Update API
kyc-check-requested	KYCService	Calls external KYC API → updates PostgreSQL kyc_result table	SET kyc:paragshah '{"status":"completed","result":"pass"}'
credit-score-requested	CreditScoreService	Calls credit bureau → inserts into credit_score_result table	SET credit:paragshah '{"score":780,"status":"completed"}'
fraud-check-requested	FraudDetectionService	Runs ML Model → inserts into fraud_result table	SET fraud:paragshah '{"risk":"low","status":"completed"}'

All three publish result events:

• kyc-completed, credit-score-completed, fraud-check-completed

• LoanEvaluationService listens to all three and waits until all results are in (fan-in pattern).

Step 5 – Loan Evaluation & Offer

• Once all three results available, LoanEvaluationService calculates eligibility:

  • Updates loan_application table → status = OFFER_GENERATED

  • Publishes Kafka event: loan-offer-generated

• Consumers:

  • CacheUpdaterConsumer updates cache:SET loan:123:status '{"status":"offer_generated","offer_amount":500000}'

  • CosmosDBWriterConsumer stores event for dashboard history.

Step 6 – Loan Agreement & e-Sign

• Parag reviews offer → clicks "Accept"

• LoanAgreementService generates PDF in Azure Blob → e-Sign

• Updates PostgreSQL table loan_agreement → signed = true

• Publishes Kafka event: loan-agreement-signed

• Cache:SET loan:123:agreement '{"status":"signed"}'

Step 7 – Disbursement

• DisbursementService triggered → Updates disbursement table

• Publishes Kafka event: loan-disbursed

• Cache:SET loan:123:status '{"status":"disbursed","amount":500000}'

• CosmosDB: Final event inserted → used for dashboard analytics.

📊 Data Storage Across Layers

Component	Sample Data for Parag Shah
PostgreSQL BDR	Tables: customer_profile, loan_application, kyc_result, credit_score_result, fraud_result, loan_agreement, disbursement
CosmosDB	loan_events container with JSON docs for each event (loan-application-initiated, kyc-completed, credit-score-completed, etc.)
Redis Cache	Keys: session:paragshah, loan:123:status, kyc:paragshah, credit:paragshah, fraud:paragshah → all stored dynamically
Kafka	Topics: loan-application-initiated, document-uploaded, kyc-check-requested, kyc-completed, credit-score-requested, credit-score-completed, fraud-check-requested, fraud-check-completed, loan-offer-generated, loan-agreement-signed, loan-disbursed

⚡ Azure Components

• API Gateway → Azure API Management

• Orchestration → Azure Logic Apps or Durable Functions

• Compute → AKS for all microservices

• Data → Azure Database for PostgreSQL (BDR enabled), CosmosDB (NoSQL), Azure Redis Cache, Azure Blob

• Event Streaming → Azure Event Hubs (Kafka interface)

• Monitoring → Azure Monitor, Application Insights, Grafana