Caching
Cache Invalidation
IntermediateWhy cache invalidation is hard, how stale data propagates, how invalidation failures cause cache poisoning, and how to design invalidation-safe patterns.
Step 1 of 5
Warm Cache: High Hit Rate
A warm cache serves reads from Redis, absorbing 90-99% of read requests before they reach PostgreSQL. Each cache entry has a TTL that determines its maximum staleness window. Under steady operation, the hit rate is high and database load is low.
user:123
valid
Alice: Premium
TTL 240s
user:456
valid
Bob: Free
TTL 180s
product:789
valid
$49.99
TTL 300s
product:101
valid
In Stock
TTL 90s
feed:top10
valid
[post1, post2, ...]
TTL 60s
Hit Rate
96%
Miss Rate
4%
Cache QPS
3,200
DB QPS
120
Warm cache: 96% hit rate, DB at 120 QPS
Key Takeaways
- →High cache hit rate keeps database load proportional to new/unique requests only
- →TTL controls maximum staleness window: shorter TTL = fresher data, more DB load
1 / 5