Caching Strategy Selector — caching strategy optional guide

3. Write Behind (Write Back)

: App → Cache  → asynchronous as DB 
: App → Cache confirmation → upper 

strategy optional

cache invalid-ize pattern

1. TTL (Time-To-Live)

# TTL configuration guide
CACHE_TTL = {
    "user_profile": 300,      # 5minutes — week change 
    "product_list": 60,       # 1minutes — between change frequency
    "stock_count": 10,        # 10seconds — week change
    "config": 3600,           # 1between — of change  
    "session": 86400,         # 24between — session countpeople
}

# Jitter addition (cache  )
import random
ttl = base_ttl + random.randint(0, base_ttl // 10)

2. Event-Based Invalidation

# event  invalid-ize
@event_handler("user.updated")
def invalidate_user_cache(event):
    cache.delete(f"user:{event.user_id}")
    cache.delete(f"user_profile:{event.user_id}")
    # related  cachealso invalid-ize
    cache.delete("user_list:page:*")

3. Version-Based

# before keyas in invalid-ize
version = cache.get("product_version") or 1
key = f"product_list:v{version}"
# invalid-ize: beforeonly 
cache.incr("product_version")

cache resolution

Cache Stampede (Thundering Herd)

: cache only   requestthis  DBin