Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
affaan-m avatar

Redis Patterns

  • 1.4k installs
  • 238k repo stars
  • Updated August 5, 2026
  • affaan-m/ecc

This is a copy of redis-patterns by affaan-m - installs and ranking accrue to the original listing.

redis-patterns is a Claude Code reference skill that supplies battle-tested Redis implementations for caching, rate limiting, distributed locks, sessions, and pub/sub so backend developers avoid reinventing production-gr

About

redis-patterns is an ECC backend reference skill covering Redis data structures, atomic workflows, and production configuration for common application workloads. The skill explains when to use strings, hashes, lists, sets, sorted sets, and streams, and how multi-step operations require Lua scripts, MULTI/EXEC transactions, or explicit synchronization beyond single-command atomicity. Developers reach for redis-patterns when adding caching layers, rate limiters, distributed locks, session stores, or pub/sub and Redis Streams messaging, including pooling, eviction, and clustering setup. The included data-structure cheat sheet maps use cases to Redis types with concrete command examples for backend services.

  • Complete Redis data structure cheat sheet with 7 core patterns
  • Production caching strategies and eviction policies
  • Distributed locks, rate limiting, and atomic operations guidance
  • Connection pooling, clustering, and persistence recommendations
  • Pub/Sub, Streams, and messaging architecture patterns

Redis Patterns by the numbers

  • 1,398 all-time installs (skills.sh)
  • +83 installs in the week ending Aug 5, 2026 (Skillselion tracking)
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/affaan-m/ecc --skill redis-patterns

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs1.4k
repo stars238k
Last updatedAugust 5, 2026
Repositoryaffaan-m/ecc

What Redis patterns handle caching and distributed locks?

Get battle-tested Redis patterns for caching, rate limiting, distributed locks, sessions, and pub/sub without reinventing production-grade implementations.

Who is it for?

Backend engineers adding Redis caching, throttling, locks, or pub/sub to APIs and web services.

Skip if: Projects needing only exact key-value lookups without in-memory caching or coordination requirements.

When should I use this skill?

A developer asks to add Redis caching, rate limits, distributed locks, sessions, or pub/sub to a backend service.

What you get

Production Redis configurations with chosen data structures, atomic scripts, connection pools, and messaging patterns.

  • Redis command patterns
  • Atomic Lua scripts
  • Production connection pool config

By the numbers

  • Documents Redis strings, hashes, lists, sets, sorted sets, and streams data structures
  • Includes a data-structure cheat sheet mapping use cases to Redis types

Files

SKILL.mdMarkdownGitHub ↗

Redis Patterns

一般的なバックエンド使用例に対するRedisベストプラクティスの参考資料。

How It Works

Redisはメモリ内データ構造ストアで、文字列、ハッシュ、リスト、セット、ソート済みセット、ストリームなどをサポートします。単一インスタンスでは個々のRedisコマンドは原子的ですが、マルチステップワークフローはLuaスクリプト、MULTI/EXECトランザクション、または明示的な同期化が必要です。RDBスナップショットまたはAOFログを通じてデータをオプションで永続化します。クライアントはRESPプロトコルを使用してTCP経由で通信します。接続プール不可欠でリクエストごとのハンドシェイクオーバーヘッドを回避します。

When to Activate

  • アプリケーションにキャッシング追加
  • レート制限またはスロットリング実装
  • 分散ロックまたはコーディネーション構築
  • セッションまたはトークンストレージ設定
  • Pub/SubまたはRedis Streams for messaging使用
  • 本番環境でRedis設定(プール、削除、クラスタリング)

Data Structure Cheat Sheet

Use CaseStructureExample Key
Simple cacheStringproduct:123
User sessionHashsession:abc
LeaderboardSorted Setscores:weekly
Unique visitorsSetvisitors:2024-01-01
Activity feedListfeed:user:456
Event streamStreamevents:orders
Counters / rate limitsString (INCR)ratelimit:user:123
Bloom filter / HLLHyperLogLoghll:pageviews

Core Patterns

Cache-Aside (Lazy Loading)

import redis
import json

r = redis.Redis(host='localhost', port=6379, decode_responses=True)

def get_product(product_id: int):
    cache_key = f"product:{product_id}"
    cached = r.get(cache_key)

    if cached:
        return json.loads(cached)

    product = db.query("SELECT * FROM products WHERE id = %s", product_id)
    r.setex(cache_key, 3600, json.dumps(product))  # TTL: 1 hour
    return product

Write-Through Cache

def update_product(product_id: int, data: dict):
    # DB書き込み先
    db.execute("UPDATE products SET ... WHERE id = %s", product_id)

    # キャッシュを即座に更新
    cache_key = f"product:{product_id}"
    r.setex(cache_key, 3600, json.dumps(data))

Cache Invalidation

# タグベース削除 — セット内で関連キーをグループ化
def cache_product(product_id: int, category_id: int, data: dict):
    key = f"product:{product_id}"
    tag = f"tag:category:{category_id}"
    pipe = r.pipeline(transaction=True)
    pipe.setex(key, 3600, json.dumps(data))
    pipe.sadd(tag, key)
    pipe.expire(tag, 3600)
    pipe.execute()

def invalidate_category(category_id: int):
    tag = f"tag:category:{category_id}"
    keys = r.smembers(tag)
    if keys:
        r.delete(*keys)
    r.delete(tag)

Session Storage

import time
import uuid

def create_session(user_id: int, ttl: int = 86400) -> str:
    session_id = str(uuid.uuid4())
    key = f"session:{session_id}"
    pipe = r.pipeline(transaction=True)
    pipe.hset(key, mapping={
        "user_id": user_id,
        "created_at": int(time.time()),
    })
    pipe.expire(key, ttl)
    pipe.execute()
    return session_id

def get_session(session_id: str) -> dict | None:
    data = r.hgetall(f"session:{session_id}")
    return data if data else None

def delete_session(session_id: str):
    r.delete(f"session:{session_id}")

Rate Limiting

Fixed Window (Simple)

def is_rate_limited(user_id: int, limit: int = 100, window: int = 60) -> bool:
    key = f"ratelimit:{user_id}:{int(time.time()) // window}"
    pipe = r.pipeline(transaction=True)
    pipe.incr(key)
    pipe.expire(key, window)
    count, _ = pipe.execute()
    return count > limit

Sliding Window (Lua — Atomic)

-- sliding_window.lua
local key = KEYS[1]
local now = tonumber(ARGV[1])
local window = tonumber(ARGV[2])
local limit = tonumber(ARGV[3])

redis.call('ZREMRANGEBYSCORE', key, 0, now - window)
local count = redis.call('ZCARD', key)

if count < limit then
    -- Use unique member (now + sequence) to avoid collisions within the same millisecond
    local seq_key = key .. ':seq'
    local seq = redis.call('INCR', seq_key)
    redis.call('EXPIRE', seq_key, math.ceil(window / 1000))
    redis.call('ZADD', key, now, now .. '-' .. seq)
    redis.call('EXPIRE', key, math.ceil(window / 1000))
    return 1
end
return 0
sliding_window = r.register_script(open('sliding_window.lua').read())

def allow_request(user_id: int) -> bool:
    key = f"ratelimit:sliding:{user_id}"
    now = int(time.time() * 1000)
    return bool(sliding_window(keys=[key], args=[now, 60000, 100]))

Distributed Locks

Distributed Lock (Single Node — SET NX PX)

import uuid

def acquire_lock(resource: str, ttl_ms: int = 5000) -> str | None:
    lock_key = f"lock:{resource}"
    token = str(uuid.uuid4())
    acquired = r.set(lock_key, token, px=ttl_ms, nx=True)
    return token if acquired else None

def release_lock(resource: str, token: str) -> bool:
    release_script = """
    if redis.call('get', KEYS[1]) == ARGV[1] then
        return redis.call('del', KEYS[1])
    else
        return 0
    end
    """
    result = r.eval(release_script, 1, f"lock:{resource}", token)
    return bool(result)

# Usage
token = acquire_lock("order:payment:123")
if token:
    try:
        process_payment()
    finally:
        release_lock("order:payment:123", token)
マルチノード設定の場合、フルRedlockアルゴリズムを実装する redlock-py ライブラリを使用してください。

Pub/Sub & Streams

Pub/Sub (Fire-and-Forget)

# Publisher
def publish_event(channel: str, payload: dict):
    r.publish(channel, json.dumps(payload))

# Subscriber (blocking — run in separate thread/process)
def subscribe_events(channel: str):
    pubsub = r.pubsub()
    pubsub.subscribe(channel)
    for message in pubsub.listen():
        if message['type'] == 'message':
            handle(json.loads(message['data']))

Redis Streams (Durable Queue)

# Producer
def emit(stream: str, event: dict):
    r.xadd(stream, event, maxlen=10000)  # Cap stream length

# Consumer group — guarantees at-least-once delivery
try:
    r.xgroup_create('events:orders', 'processor', id='0', mkstream=True)
except Exception:
    pass  # Group already exists

def consume(stream: str, group: str, consumer: str):
    while True:
        messages = r.xreadgroup(group, consumer, {stream: '>'}, count=10, block=2000)
        for _, entries in (messages or []):
            for msg_id, data in entries:
                process(data)
                r.xack(stream, group, msg_id)
配信保証、コンシューマーグループ、または再生が必要な場合、Pub/Sub代わりにStreamsを優先してください。

Key Design

Naming Conventions

# Pattern: resource:id:field
user:123:profile
order:456:status
cache:product:789

# Pattern: namespace:resource:id
myapp:session:abc123
myapp:ratelimit:user:123

# Pattern: resource:date (time-bound keys)
stats:pageviews:2024-01-01

TTL Strategy

Data TypeSuggested TTL
User session24h (86400)
API response cache5–15 min
Rate limit windowMatch window size
Short-lived tokens5–10 min
Leaderboard1h–24h
Static/reference data1h–1 week

常にTTLを設定してください。TTLなしのキーは無限に蓄積してメモリ圧力を引き起こします。

Connection Management

Connection Pooling

from redis import ConnectionPool, Redis

pool = ConnectionPool(
    host='localhost',
    port=6379,
    db=0,
    max_connections=20,
    decode_responses=True,
    socket_connect_timeout=2,
    socket_timeout=2,
)

r = Redis(connection_pool=pool)

Cluster Mode

from redis.cluster import RedisCluster

r = RedisCluster(
    startup_nodes=[{"host": "redis-1", "port": 6379}],
    decode_responses=True,
    skip_full_coverage_check=True,
)

Sentinel (High Availability)

from redis.sentinel import Sentinel

sentinel = Sentinel(
    [('sentinel-1', 26379), ('sentinel-2', 26379)],
    socket_timeout=0.5,
)
master = sentinel.master_for('mymaster', decode_responses=True)
replica = sentinel.slave_for('mymaster', decode_responses=True)

Eviction Policies

PolicyBehaviorBest For
noevictionError on write when fullQueues / critical data
allkeys-lruEvict least recently usedGeneral cache
volatile-lruLRU only among keys with TTLMixed data store
allkeys-lfuEvict least frequently usedSkewed access patterns
volatile-ttlEvict soonest-to-expirePrioritize long-lived data

redis.confを通じて設定:maxmemory-policy allkeys-lru

Anti-Patterns

Anti-PatternProblemFix
Keys with no TTLMemory grows unboundedAlways set TTL
KEYS * in productionBlocks the server (O(N))Use SCAN cursor
Storing large blobs (>100KB)Slow serialization, memory pressureStore reference + fetch from object store
Single Redis for everythingNo isolation between cache & queueUse separate DBs or instances
Ignoring connection pool limitsConnection exhaustion under loadSize pool to workload
Not handling cache miss stampedeThundering herd on cold startUse locks or probabilistic early expiry
FLUSHALL without thoughtWipes entire instanceScope deletes by key pattern

Cache Miss Stampede Prevention

import threading

_locks: dict[str, threading.Lock] = {}
_locks_mutex = threading.Lock()

def get_with_lock(key: str, fetch_fn, ttl: int = 300):
    cached = r.get(key)
    if cached:
        return json.loads(cached)

    with _locks_mutex:
        if key not in _locks:
            _locks[key] = threading.Lock()
        lock = _locks[key]
    with lock:
        cached = r.get(key)  # Re-check after acquiring lock
        if cached:
            return json.loads(cached)
        value = fetch_fn()
        r.setex(key, ttl, json.dumps(value))
        return value
マルチプロセスデプロイメント:インプロセスロックを上記の分散ロックセクション から acquire_lock/release_lock に置き換えてください。

Examples

Django/Flask APIエンドポイントにキャッシング追加: レスポンスに5分TTLでCache-asideを使用。リクエストパラメータでキーを指定。

ユーザーごとにAPIレート制限: 低トラフィックエンドポイントに固定ウィンドウを pipeline(transaction=True) で使用;正確なユーザーごと制限にはsliding-windowの Lua使用。

ワーカー間のバックグラウンドジョブ調整: 予想ジョブ期間を超えるTTLで acquire_lock を使用。常に finally ブロックでリリース。

複数購読者への通知のファンアウト: ファイアアンドフォーゲットにPub/Subを使用。保証配信または再生が必要な場合、Streamsに切り替え。

Quick Reference

PatternWhen to Use
Cache-asideRead-heavy, tolerate slight staleness
Write-throughStrong consistency required
Distributed lockPrevent concurrent access to a resource
Sliding window rate limitAccurate per-user throttling
Redis StreamsDurable event queue with consumer groups
Pub/SubBroadcast with no delivery guarantees needed
Sorted Set leaderboardRanked scoring, pagination
HyperLogLogApproximate unique count at low memory

Related

  • Skill: postgres-patterns — リレーショナルデータパターン
  • Skill: backend-patterns — APIおよびサービスレイヤーパターン
  • Skill: database-migrations — スキーマバージョニング
  • Skill: django-patterns — Djangoキャッシュフレームワーク統合
  • Agent: database-reviewer — 全データベースレビューワークフロー

Related skills

How it compares

Pick redis-patterns over generic database skills when the workload is in-memory coordination, caching, or messaging on Redis specifically.

FAQ

What Redis use cases does redis-patterns cover?

redis-patterns covers caching, rate limiting, distributed locks, session and token storage, pub/sub, and Redis Streams messaging. The skill also addresses connection pooling, eviction policies, and clustering for production deployments.

How does redis-patterns handle multi-step Redis operations?

redis-patterns explains that individual Redis commands are atomic on one instance, but multi-step workflows need Lua scripts, MULTI/EXEC transactions, or explicit synchronization. The skill recommends patterns that avoid race conditions in locks and counters.

Backend & APIsbackendintegrations

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.