arturAtDeliveroo

# PR #1136 - CRITICAL Performance Issues

## Problem Statement
- Connections increased 4-5x
- Hit rate only 65% (expected: 95%+)
- Cache latency 80ms (expected: microseconds)

## Root Causes Found

### 🔥 CRITICAL #1: Async Cache Update (lines 355-358)

arturAtDeliveroo

--- a/pkg/redisconn/generic_swr_cache.go
+++ b/pkg/redisconn/generic_swr_cache.go
@@ -213,10 +213,9 @@ func (c *GenericSWRCache) GetBulk(keys []string) (results map[string]*GenericCac
 		}

 		// Check if entry is invalid (> InvalidAfterSeconds)
 		if !entry.IsValid(invalidAfter) {
 			// Invalid entries are treated as missing - must fetch from Redis
 			missing = append(missing, key)
-			misses++

arturAtDeliveroo

// File: pkg/redisconn/generic_swr_cache.go
// Surgical change: Add memory breakdown metrics by cache state

// In reportDataFreshness method (around line 808), REPLACE entire function:
func (c *GenericSWRCache) reportDataFreshness() {
	now := time.Now()
	staleThreshold := time.Duration(c.config.StaleAfterSeconds) * time.Second
	invalidThreshold := time.Duration(c.config.InvalidAfterSeconds) * time.Second
	zombieThreshold := 5 * time.Minute

arturAtDeliveroo

// File: app/config/config.go
// Changes: Add InvalidAfterSeconds field, adjust StaleAfterSeconds default

type GenericCache struct {
	MaxMemoryMB int `envconfig:"CACHE_MAX_MEMORY_MB" default:"300"`

	// StaleAfterSeconds: Time after which cached data is stale (served instantly, queued for refresh)
	// Default: 5 seconds
	StaleAfterSeconds int `envconfig:"CACHE_STALE_AFTER_SECONDS" default:"5"`