feat(storage): fix critical socket exhaustion during metadata reading

CRITICAL ISSUE FIXED: Service initialization was failing due to socket exhaustion
when reading metadata for 1400+ items during index rebuild

Changes:
- Add batch metadata reading to prevent 2000+ concurrent requests
- Implement strict concurrency control (3 max concurrent requests)
- Add proper yielding between batches to prevent event loop blocking
- Reduce batch sizes during initialization (50 → 25 items per batch)
- Add getMetadataBatch() and getVerbMetadataBatch() methods to S3 storage
- Update StorageAdapter interface with batch methods
- Add production environment auto-detection for smart logging
- Auto-cleanup legacy /index folder during initialization

Socket usage: Reduced from 1400+ concurrent to 3 max concurrent

Expected production result:
- Service initialization will complete successfully
- firehoseServiceInitialized: true
- Data collection will begin normally
- No more socket exhaustion errors (100 socket limit exceeded)

Fixes: #socket-exhaustion
Breaking: None - backward compatible with fallback modes
This commit is contained in:
David Snelling 2025-08-07 11:46:07 -07:00
parent 6648149e90
commit 1b265e8c42
7 changed files with 485 additions and 52 deletions

View file

@ -84,3 +84,103 @@ export function isThreadingAvailable(): boolean {
export async function isThreadingAvailableAsync(): Promise<boolean> {
return areWebWorkersAvailable() || (await areWorkerThreadsAvailable())
}
/**
* Auto-detect production environment to minimize logging costs
*/
export function isProductionEnvironment(): boolean {
// Node.js environment detection
if (isNode()) {
// Check common production environment indicators
const nodeEnv = process.env.NODE_ENV?.toLowerCase()
if (nodeEnv === 'production' || nodeEnv === 'prod') return true
// Google Cloud Run detection
if (process.env.K_SERVICE || process.env.GOOGLE_CLOUD_PROJECT) return true
// AWS Lambda detection
if (process.env.AWS_LAMBDA_FUNCTION_NAME || process.env.AWS_EXECUTION_ENV) return true
// Azure Functions detection
if (process.env.AZURE_FUNCTIONS_ENVIRONMENT || process.env.WEBSITE_SITE_NAME) return true
// Vercel detection
if (process.env.VERCEL || process.env.VERCEL_ENV === 'production') return true
// Netlify detection
if (process.env.NETLIFY && process.env.CONTEXT === 'production') return true
// Heroku detection
if (process.env.DYNO && process.env.NODE_ENV !== 'development') return true
// Railway detection
if (process.env.RAILWAY_ENVIRONMENT === 'production') return true
// Fly.io detection
if (process.env.FLY_APP_NAME && process.env.FLY_REGION) return true
// Docker in production (common patterns)
if (process.env.DOCKER_ENV === 'production' || process.env.ENVIRONMENT === 'production') return true
// Generic production indicators
if (process.env.PROD === 'true' || process.env.PRODUCTION === 'true') return true
}
// Browser environment - assume development unless explicitly production
if (isBrowser()) {
// Check for production domain patterns
const hostname = window?.location?.hostname
if (hostname) {
// Avoid logging on production domains
if (hostname.includes('.com') || hostname.includes('.org') || hostname.includes('.net')) {
return !hostname.includes('localhost') && !hostname.includes('127.0.0.1') && !hostname.includes('dev')
}
}
}
return false
}
/**
* Get appropriate log level based on environment
*/
export function getLogLevel(): 'silent' | 'error' | 'warn' | 'info' | 'verbose' {
// Explicit log level override
const explicitLevel = process.env.BRAINY_LOG_LEVEL?.toLowerCase()
if (explicitLevel && ['silent', 'error', 'warn', 'info', 'verbose'].includes(explicitLevel)) {
return explicitLevel as 'silent' | 'error' | 'warn' | 'info' | 'verbose'
}
// Auto-detect based on environment
if (isProductionEnvironment()) {
return 'error' // Only log errors in production to minimize costs
}
// Development environments get more verbose logging
if (process.env.NODE_ENV === 'development' || process.env.NODE_ENV === 'dev') {
return 'verbose'
}
// Test environments should be quieter
if (process.env.NODE_ENV === 'test') {
return 'warn'
}
// Default to info level
return 'info'
}
/**
* Check if logging should be enabled for a given level
*/
export function shouldLog(level: 'error' | 'warn' | 'info' | 'verbose'): boolean {
const currentLevel = getLogLevel()
if (currentLevel === 'silent') return false
const levels = ['error', 'warn', 'info', 'verbose']
const currentIndex = levels.indexOf(currentLevel)
const messageIndex = levels.indexOf(level)
return messageIndex <= currentIndex
}