salesforce-reliability-patterns

# Salesforce Reliability Patterns ## Overview Production-grade reliability patterns for Salesforce integrations: circuit breakers for API outages, idempotent operations using External IDs, graceful degradation with cached data, and dead letter queues for failed operations. ## Prerequisites - jsforce connection configured - Understanding of Salesforce error codes (see `salesforce-common-errors`) - Redis or database for state management (optional) - opossum or similar circuit breaker library ## Instructions ### Step 1: Circuit Breaker for Salesforce API ```typescript import CircuitBreaker from 'opossum'; import { getConnection } from './salesforce/connection'; // Circuit breaker wraps all Salesforce calls const sfBreaker = new CircuitBreaker( async (fn: () => Promise<any>) => fn(), { timeout: 30000, // SF calls can be slow — 30s timeout errorThresholdPercentage: 50, // Open circuit at 50% error rate resetTimeout: 60000, // Try again after 1 minute volumeThreshold: 10, // Need 10 calls before evaluating errorFilter: (error: any) => { // Don't count client errors as circuit-breaking failures const nonCircuitErrors = ['INVALID_FIELD', 'MALFORMED_QUERY', 'REQUIRED_FIELD_MISSING']; return nonCircuitErrors.includes(error.errorCode); }, } ); sfBreaker.on('open', () => { console.error('CIRCUIT OPEN: Salesforce API failing — requests will fail fast'); // Alert ops team }); sfBreaker.on('halfO...