Session Cache with TTL
Build login session storage using SET with EX, GET, and DEL. Add a check that denies access when TTL is expired.
Data / Knowledge library
Redis
in-memory key-value store for caching and real-time data
Why learn Redis?
Redis provides very fast read/write access for data that must be served in milliseconds, such as caches, sessions, counters, queues, rate limits, and real-time leaderboard/state workloads.
Where it came from
Redis was created by Salvatore Sanfilippo and first released in 2009. The name Redis comes from 'REmote DIctionary Server'. Redis Cluster (hash-slot based sharding) arrived in Redis 3.0. Sentinel became the standard built-in option for Redis high availability/failover in OSS deployments.
- Redis was created by Salvatore Sanfilippo and first released in 2009.
- The name Redis comes from 'REmote DIctionary Server'.
- Redis Cluster (hash-slot based sharding) arrived in Redis 3.0.
- Sentinel became the standard built-in option for Redis high availability/failover in OSS deployments.
Build your understanding
Work through these exercises in order. Try the task before opening a reference.
- Build login session storage using SET with EX, GET, and DEL. Add a check that denies access when TTL is expired.
- Implement a fixed-window limiter using INCR + EXPIRE per IP. Return remaining requests and reset time.
- Use ZADD, ZINCRBY, ZREVRANGE, and periodic persistence settings to support top-N ranking with fast reads.
- Create a queue with BRPOPLPUSH (or equivalent pattern), retry handling, and dead-letter tracking.
- Use WATCH/MULTI/EXEC or Lua to prevent overselling when multiple buyers hit the same stock key.
More ways to practice
Per-IP Rate Limiter
Implement a fixed-window limiter using INCR + EXPIRE per IP. Return remaining requests and reset time.
Realtime Leaderboard
Use ZADD, ZINCRBY, ZREVRANGE, and periodic persistence settings to support top-N ranking with fast reads.
Reliable Worker Queue
Create a queue with BRPOPLPUSH (or equivalent pattern), retry handling, and dead-letter tracking.
Atomic Inventory Update
Use WATCH/MULTI/EXEC or Lua to prevent overselling when multiple buyers hit the same stock key.
Documentation and sources
- Official Redis docs (redis.io)
- Command reference (redis.io)
- Data types overview (redis.io)
- Persistence (RDB and AOF) (redis.io)
Catalog research notes
- Redis 7.x expanded server-side programmability (Redis Functions) and sharded Pub/Sub improvements.
- RESP3 adoption has increased across modern Redis clients for richer protocol semantics.
- Redis Stack workflows (JSON, Search, TimeSeries, Bloom) became common for app-level data features, including vector search.
- 2024 licensing changes for Redis source code triggered ecosystem shifts, including the Valkey fork.
Connect the ideas
Alternatives
Explore next
ValkeyMemcachedPostgreSQLApache Kafka (kafka)RabbitMQApache Airflow (airflow)Apache CassandraDjangoRuby on Rails (Rails)
Real World Example
Real-time Inventory Caching for E-commerce
Online retailers managing high-traffic flash sales.
The business problem
Database contention occurs when thousands of concurrent users query the primary relational database for item availability, leading to latency spikes and potential checkout failures.
How this topic solves it
Implement Redis Hashes to store inventory counts in-memory, allowing the application to check stock levels with sub-millisecond latency before hitting the primary database.
- Initialize a Redis Hash key for each product ID to store fields like 'stock_count' and 'reserved_count'.
- On product page load, query the Redis Hash to display real-time availability.
- During checkout, use atomic operations to decrement the stock count within the Redis Hash.
- Periodically synchronize the final transaction state from Redis back to the primary relational database.
Shows how user requests interact with Redis to offload the primary database.
Customer browsing items
Application logic
In-memory inventory state
Persistent record of truth
- User → Request stock → App Server
- App Server → Read/Update stock → Redis Hash
- Redis Hash → Async sync → Primary DB
Implementation example
Uses Redis Hashes to manage inventory fields and atomic increments to ensure thread-safe stock updates.
HSET product:101 stock 50 reserved 0 // Initialize product inventory
HINCRBY product:101 stock -1 // Atomically decrement stock during purchaseExpected business outcome
Reduced load on the primary database and a smoother, more responsive user experience during high-traffic events.
Limitations and checks
Redis is an in-memory store, meaning data persistence depends on configured snapshotting or append-only file settings; if the server crashes before a sync, updates in memory might be lost. Check persistence status using the INFO command.