Redis Example

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122

import redis
 
# Connect to Redis
r = redis.Redis(host='localhost', port=6379, db=0, decode_responses=True)
 
# ========================================
# STRINGS: Basic key-value pairs
# ========================================
 
# Simple string
r.set('user:1001:name', 'Alice Johnson')
name = r.get('user:1001:name')  # 'Alice Johnson'
 
# String with expiration (TTL)
r.setex('session:abc123', 3600, 'user_data_here')  # Expires in 1 hour
 
# Atomic increment (counters)
r.set('page_views', 0)
r.incr('page_views')      # Returns 1
r.incrby('page_views', 5)  # Returns 6
 
# Set only if not exists (distributed lock primitive)
acquired = r.setnx('lock:resource', 'owner_id')  # Returns True if set
 
# Set with expiration only if key doesn't exist (better lock)
r.set('lock:resource', 'owner_id', nx=True, ex=30)
 
# ========================================
# LISTS: Ordered collections
# ========================================
 
# Add to lists
r.lpush('notifications:1001', 'msg3', 'msg2', 'msg1')  # Push to left
r.rpush('queue:tasks', 'task1', 'task2')  # Push to right
 
# Pop from lists
task = r.lpop('queue:tasks')  # Returns 'task1'
task = r.brpop('queue:tasks', timeout=5)  # Blocking pop, waits 5s
 
# Range access
notifications = r.lrange('notifications:1001', 0, 9)  # First 10
 
# Trim list (keep only recent)
r.ltrim('notifications:1001', 0, 99)  # Keep only 100 most recent
 
# ========================================
# SETS: Unique unordered collections
# ========================================
 
# Add members
r.sadd('user:1001:tags', 'premium', 'developer', 'sf-bay')
r.sadd('user:1002:tags', 'free', 'designer', 'sf-bay')
 
# Check membership
is_premium = r.sismember('user:1001:tags', 'premium')  # True
 
# Set operations
common_tags = r.sinter('user:1001:tags', 'user:1002:tags')  # {'sf-bay'}
all_tags = r.sunion('user:1001:tags', 'user:1002:tags')
 
# Random member (for sampling)
random_tag = r.srandmember('user:1001:tags')
 
# ========================================
# SORTED SETS: Ordered by score
# ========================================
 
# Add with scores (leaderboard)
r.zadd('leaderboard:game1', {'alice': 1500, 'bob': 1200, 'charlie': 1800})
 
# Increment score
r.zincrby('leaderboard:game1', 100, 'alice')  # Alice now 1600
 
# Rank queries
rank = r.zrevrank('leaderboard:game1', 'alice')  # 1 (0-indexed, descending)
top_3 = r.zrevrange('leaderboard:game1', 0, 2, withscores=True)
# [('charlie', 1800), ('alice', 1600), ('bob', 1200)]
 
# Score range (e.g., recent time-based events)
# Using timestamp as score
now = time.time()
r.zadd('events:user:1001', {f'event_{i}': now + i for i in range(10)})
recent = r.zrangebyscore('events:user:1001', now - 3600, now)  # Last hour
 
# ========================================
# HASHES: Field-value maps
# ========================================
 
# Store object as hash
r.hset('user:1001', mapping={
    'email': 'alice@example.com',
    'name': 'Alice Johnson',
    'login_count': 0,
    'last_login': '',
})
 
# Get single field
email = r.hget('user:1001', 'email')
 
# Get multiple fields
user_data = r.hmget('user:1001', ['email', 'name'])
 
# Get all fields
user = r.hgetall('user:1001')
 
# Increment field
r.hincrby('user:1001', 'login_count', 1)
 
# ========================================
# HYPERLOGLOG: Cardinality estimation
# ========================================
 
# Count unique visitors (O(1) memory regardless of count)
r.pfadd('unique_visitors:2024-01-15', 'user1', 'user2', 'user3')
r.pfadd('unique_visitors:2024-01-15', 'user2', 'user4')  # user2 not re-counted
 
count = r.pfcount('unique_visitors:2024-01-15')  # ~4 (approx, <1% error)
 
# Merge multiple HyperLogLogs
r.pfmerge('unique_visitors:week', 
          'unique_visitors:2024-01-15', 
          'unique_visitors:2024-01-16')

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90

class CacheAsidePattern:
    """
    Cache-Aside pattern implementation.
    Also known as lazy-loading or look-aside cache.
    """
    
    def __init__(self, redis_client, database, default_ttl=3600):
        self.redis = redis_client
        self.db = database
        self.default_ttl = default_ttl
    
    def get_user(self, user_id: str) -> dict:
        """
        Get user with cache-aside pattern.
        
        1. Check cache first (fast path)
        2. On miss, load from database
        3. Populate cache for next request
        """
        cache_key = f"cache:user:{user_id}"
        
        # Step 1: Try cache
        cached = self.redis.get(cache_key)
        if cached:
            return json.loads(cached)
        
        # Step 2: Cache miss - load from database
        user = self.db.get_user(user_id)
        if user is None:
            return None
        
        # Step 3: Populate cache with TTL
        self.redis.setex(
            cache_key, 
            self.default_ttl, 
            json.dumps(user)
        )
        
        return user
    
    def update_user(self, user_id: str, data: dict) -> None:
        """
        Update user with cache invalidation.
        
        Strategy: Write to database, then invalidate cache.
        Do NOT write to cache directly (data inconsistency risk).
        """
        # Update database (source of truth)
        self.db.update_user(user_id, data)
        
        # Invalidate cache - next read will reload
        cache_key = f"cache:user:{user_id}"
        self.redis.delete(cache_key)
    
    def delete_user(self, user_id: str) -> None:
        """Delete user and invalidate cache."""
        self.db.delete_user(user_id)
        self.redis.delete(f"cache:user:{user_id}")
 
 
# ========================================
# Pattern 2: Write-Through Cache
# ========================================
 
class WriteThroughPattern:
    """
    Write-Through: Write to cache AND database synchronously.
    Ensures cache is always up-to-date.
    """
    
    def update_user(self, user_id: str, data: dict) -> None:
        """
        Update database and cache synchronously.
        
        Both writes must succeed. Consider:
        - What if cache write fails after DB write?
        - What if DB write fails after cache write?
        """
        cache_key = f"cache:user:{user_id}"
        
        # Write to database first
        self.db.update_user(user_id, data)
        
        # Then write to cache with TTL
        user = self.db.get_user(user_id)  # Get complete fresh state
        self.redis.setex(
            cache_key,
            self.default_ttl,
            json.dumps(user)
        )

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123

import uuid
import time
 
class DistributedLock:
    """
    Distributed lock implementation using Redis.
    
    Uses SET NX EX for atomic check-and-set with expiration.
    Lock automatically expires to prevent deadlocks.
    """
    
    def __init__(self, redis_client, lock_name: str, 
                 expire_seconds: int = 30):
        self.redis = redis_client
        self.lock_key = f"lock:{lock_name}"
        self.expire_seconds = expire_seconds
        self.lock_id = None
    
    def acquire(self, blocking: bool = True, 
                timeout: float = None) -> bool:
        """
        Acquire the lock.
        
        Args:
            blocking: If True, wait until lock is available
            timeout: Max seconds to wait (None = wait forever)
        
        Returns:
            True if lock acquired, False otherwise
        """
        self.lock_id = str(uuid.uuid4())
        start_time = time.time()
        
        while True:
            # SET key value NX EX seconds
            # NX = only if not exists, EX = expiration
            acquired = self.redis.set(
                self.lock_key,
                self.lock_id,
                nx=True,
                ex=self.expire_seconds
            )
            
            if acquired:
                return True
            
            if not blocking:
                return False
            
            # Check timeout
            if timeout is not None:
                if time.time() - start_time > timeout:
                    return False
            
            # Wait before retry (avoid spinning)
            time.sleep(0.01)
    
    def release(self) -> bool:
        """
        Release the lock safely.
        
        Uses Lua script for atomic check-and-delete.
        Only releases if we still own the lock.
        """
        if self.lock_id is None:
            return False
        
        # Lua script ensures atomic check-and-delete
        lua_script = """
        if redis.call("GET", KEYS[1]) == ARGV[1] then
            return redis.call("DEL", KEYS[1])
        else
            return 0
        end
        """
        
        result = self.redis.eval(lua_script, 1, 
                                 self.lock_key, self.lock_id)
        
        self.lock_id = None
        return result == 1
    
    def extend(self, additional_seconds: int = None) -> bool:
        """
        Extend lock expiration (for long operations).
        
        Only extends if we still own the lock.
        """
        if additional_seconds is None:
            additional_seconds = self.expire_seconds
        
        lua_script = """
        if redis.call("GET", KEYS[1]) == ARGV[1] then
            return redis.call("EXPIRE", KEYS[1], ARGV[2])
        else
            return 0
        end
        """
        
        result = self.redis.eval(
            lua_script, 1,
            self.lock_key, self.lock_id, additional_seconds
        )
        
        return result == 1
    
    def __enter__(self):
        """Context manager support."""
        if not self.acquire():
            raise Exception("Could not acquire lock")
        return self
    
    def __exit__(self, exc_type, exc_val, exc_tb):
        self.release()
        return False
 
 
# Usage
lock = DistributedLock(redis_client, "process_order:123")
 
with lock:
    # Only one process can execute this block at a time
    process_order("123")

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133

import time
 
class SlidingWindowRateLimiter:
    """
    Sliding window rate limiter using Redis sorted sets.
    
    More accurate than fixed window, prevents burst at window edges.
    """
    
    def __init__(self, redis_client, max_requests: int, 
                 window_seconds: int):
        self.redis = redis_client
        self.max_requests = max_requests
        self.window_seconds = window_seconds
    
    def is_allowed(self, client_id: str) -> tuple[bool, dict]:
        """
        Check if request is allowed and record it.
        
        Returns:
            (allowed: bool, info: dict with remaining, reset_at)
        """
        key = f"ratelimit:{client_id}"
        now = time.time()
        window_start = now - self.window_seconds
        
        # Use pipeline for atomic operation
        pipe = self.redis.pipeline()
        
        # Remove old entries outside window
        pipe.zremrangebyscore(key, 0, window_start)
        
        # Count entries in current window
        pipe.zcard(key)
        
        # Add current request (score = timestamp)
        request_id = f"{now}:{uuid.uuid4().hex[:8]}"
        pipe.zadd(key, {request_id: now})
        
        # Set expiration on key
        pipe.expire(key, self.window_seconds + 1)
        
        # Execute pipeline
        results = pipe.execute()
        
        current_count = results[1]  # zcard result
        
        allowed = current_count < self.max_requests
        remaining = max(0, self.max_requests - current_count - 1)
        
        if not allowed:
            # Remove the request we just added
            self.redis.zrem(key, request_id)
            remaining = 0
        
        # Calculate reset time (when oldest entry expires)
        oldest = self.redis.zrange(key, 0, 0, withscores=True)
        if oldest:
            reset_at = oldest[0][1] + self.window_seconds
        else:
            reset_at = now + self.window_seconds
        
        return allowed, {
            'remaining': remaining,
            'reset_at': reset_at,
            'retry_after': 0 if allowed else reset_at - now
        }
 
 
class TokenBucketRateLimiter:
    """
    Token bucket rate limiter using Redis.
    
    Allows bursts up to bucket capacity while maintaining average rate.
    """
    
    def __init__(self, redis_client, capacity: int, 
                 refill_rate: float):
        """
        Args:
            capacity: Max tokens in bucket
            refill_rate: Tokens added per second
        """
        self.redis = redis_client
        self.capacity = capacity
        self.refill_rate = refill_rate
    
    def is_allowed(self, client_id: str, tokens: int = 1) -> bool:
        """
        Try to consume tokens from the bucket.
        
        Uses Lua script for atomic token calculation.
        """
        key = f"bucket:{client_id}"
        now = time.time()
        
        lua_script = """
        local key = KEYS[1]
        local capacity = tonumber(ARGV[1])
        local refill_rate = tonumber(ARGV[2])
        local now = tonumber(ARGV[3])
        local requested = tonumber(ARGV[4])
        
        -- Get current bucket state
        local last_update = tonumber(redis.call("HGET", key, "last_update") or now)
        local tokens = tonumber(redis.call("HGET", key, "tokens") or capacity)
        
        -- Calculate tokens to add based on time elapsed
        local elapsed = now - last_update
        local new_tokens = math.min(capacity, tokens + (elapsed * refill_rate))
        
        -- Check if we have enough tokens
        if new_tokens >= requested then
            new_tokens = new_tokens - requested
            redis.call("HSET", key, "tokens", new_tokens)
            redis.call("HSET", key, "last_update", now)
            redis.call("EXPIRE", key, 86400)  -- Clean up after 1 day idle
            return 1
        else
            -- Update timestamp but don't consume
            redis.call("HSET", key, "tokens", new_tokens)
            redis.call("HSET", key, "last_update", now)
            redis.call("EXPIRE", key, 86400)
            return 0
        end
        """
        
        result = self.redis.eval(
            lua_script, 1, key,
            self.capacity, self.refill_rate, now, tokens
        )
        
        return result == 1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90

class Leaderboard:
    """
    Real-time leaderboard using Redis sorted sets.
    
    Sorted sets maintain ordering automatically, making
    leaderboard operations extremely efficient.
    """
    
    def __init__(self, redis_client, leaderboard_name: str):
        self.redis = redis_client
        self.key = f"leaderboard:{leaderboard_name}"
    
    def set_score(self, user_id: str, score: float) -> None:
        """Set/update user's score."""
        self.redis.zadd(self.key, {user_id: score})
    
    def increment_score(self, user_id: str, delta: float) -> float:
        """Atomically increment score. Returns new score."""
        return self.redis.zincrby(self.key, delta, user_id)
    
    def get_rank(self, user_id: str) -> int | None:
        """
        Get user's rank (0-indexed, highest score = rank 0).
        Returns None if user not in leaderboard.
        """
        rank = self.redis.zrevrank(self.key, user_id)
        return rank  # 0 = first place
    
    def get_score(self, user_id: str) -> float | None:
        """Get user's current score."""
        return self.redis.zscore(self.key, user_id)
    
    def get_top(self, count: int = 10) -> list[tuple[str, float]]:
        """Get top N players with scores."""
        return self.redis.zrevrange(
            self.key, 0, count - 1, withscores=True
        )
    
    def get_around(self, user_id: str, count: int = 5)         -> list[tuple[str, float]]:
        """
        Get players around the given user.
        
        Returns 'count' players above and below.
        """
        rank = self.get_rank(user_id)
        if rank is None:
            return []
        
        start = max(0, rank - count)
        end = rank + count
        
        return self.redis.zrevrange(
            self.key, start, end, withscores=True
        )
    
    def get_page(self, page: int, page_size: int = 50)         -> list[tuple[str, float]]:
        """Get paginated leaderboard results."""
        start = page * page_size
        end = start + page_size - 1
        
        return self.redis.zrevrange(
            self.key, start, end, withscores=True
        )
    
    def remove_user(self, user_id: str) -> bool:
        """Remove user from leaderboard."""
        return self.redis.zrem(self.key, user_id) == 1
    
    def get_total_count(self) -> int:
        """Total number of players in leaderboard."""
        return self.redis.zcard(self.key)
 
 
# Usage
lb = Leaderboard(redis_client, "game:puzzle:weekly")
 
# Player completes level
lb.increment_score("player123", 100)
 
# Get player's current standing
rank = lb.get_rank("player123")  # 42
score = lb.get_score("player123")  # 1500
 
# Display top 10
top_10 = lb.get_top(10)
# [('champion', 10000), ('pro_player', 9500), ...]
 
# Show players around current player
nearby = lb.get_around("player123", count=3)
# Shows 3 above and 3 below in ranking

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127

import threading
 
class RealtimeNotifications:
    """
    Real-time notification system using Redis Pub/Sub.
    
    Pub/Sub is fire-and-forget - messages are NOT persisted.
    For durable messaging, use Redis Streams instead.
    """
    
    def __init__(self, redis_client):
        self.redis = redis_client
        self.pubsub = self.redis.pubsub()
    
    def subscribe_to_user(self, user_id: str, 
                          callback: callable) -> None:
        """Subscribe to notifications for a user."""
        channel = f"notifications:{user_id}"
        self.pubsub.subscribe(**{channel: callback})
        
        # Start listening in background thread
        thread = threading.Thread(target=self.pubsub.run_in_thread, 
                                  args=(0.01,))  # 10ms interval
        thread.daemon = True
        thread.start()
    
    def notify_user(self, user_id: str, message: dict) -> int:
        """
        Send notification to a user.
        
        Returns number of subscribers who received the message.
        Note: 0 means no one is currently subscribed!
        """
        channel = f"notifications:{user_id}"
        return self.redis.publish(channel, json.dumps(message))
    
    def broadcast_to_all(self, message: dict) -> int:
        """Broadcast to all users (use with caution)."""
        return self.redis.publish("notifications:broadcast", 
                                  json.dumps(message))
    
    def subscribe_to_pattern(self, pattern: str, 
                             callback: callable) -> None:
        """
        Subscribe to channels matching a pattern.
        
        Example: "notifications:*" for all user channels
        """
        self.pubsub.psubscribe(**{pattern: callback})
 
 
# ========================================
# Redis Streams: Durable Message Streaming
# ========================================
 
class DurableEventStream:
    """
    Durable event streaming using Redis Streams.
    
    Unlike Pub/Sub, Streams PERSIST messages and support:
    - Consumer groups for distributed processing
    - Message acknowledgment
    - Reading from specific position
    - Automatic trimming by count or time
    """
    
    def __init__(self, redis_client, stream_name: str):
        self.redis = redis_client
        self.stream = stream_name
    
    def publish(self, event: dict, max_len: int = 10000) -> str:
        """
        Publish event to stream.
        
        Returns stream ID (timestamp-sequence format).
        max_len trims old entries to cap memory usage.
        """
        # XADD stream MAXLEN ~ 10000 * field value ...
        return self.redis.xadd(
            self.stream, 
            event,
            maxlen=max_len,
            approximate=True  # ~ for better performance
        )
    
    def read_latest(self, count: int = 10) -> list:
        """Read latest events without consumer group."""
        # XREVRANGE stream + - COUNT 10
        return self.redis.xrevrange(self.stream, '+', '-', count=count)
    
    def create_consumer_group(self, group_name: str, 
                              start_id: str = '0') -> bool:
        """Create consumer group for distributed processing."""
        try:
            self.redis.xgroup_create(
                self.stream, group_name, 
                id=start_id, mkstream=True
            )
            return True
        except redis.ResponseError as e:
            if "BUSYGROUP" in str(e):
                return False  # Group already exists
            raise
    
    def consume(self, group_name: str, consumer_name: str,
                count: int = 10, block_ms: int = 5000) -> list:
        """
        Consume messages as part of a consumer group.
        
        Multiple consumers can share the load - each message
        is delivered to only ONE consumer in the group.
        """
        # XREADGROUP GROUP group consumer BLOCK 5000 COUNT 10 STREAMS stream >
        result = self.redis.xreadgroup(
            group_name, consumer_name,
            {self.stream: '>'},  # '>' = only new messages
            count=count,
            block=block_ms
        )
        
        if result:
            return result[0][1]  # List of (id, data) tuples
        return []
    
    def acknowledge(self, group_name: str, message_id: str) -> int:
        """Acknowledge message processing completed."""
        return self.redis.xack(self.stream, group_name, message_id)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84

# ========================================
# RDB SNAPSHOTTING CONFIGURATION
# ========================================
 
# Save after 900 seconds if at least 1 key changed
save 900 1
# Save after 300 seconds if at least 10 keys changed
save 300 10
# Save after 60 seconds if at least 10000 keys changed
save 60 10000
 
# Disable RDB entirely
# save ""
 
# Stop accepting writes if RDB save fails
stop-writes-on-bgsave-error yes
 
# Compress RDB dump file (uses LZF, slight CPU cost)
rdbcompression yes
 
# Include CRC64 checksum in RDB file
rdbchecksum yes
 
# RDB filename
dbfilename dump.rdb
 
# Working directory (where RDB and AOF are saved)
dir /var/lib/redis
 
# ========================================
# AOF CONFIGURATION
# ========================================
 
# Enable AOF persistence
appendonly yes
 
# AOF filename
appendfilename "appendonly.aof"
 
# Sync policy options:
# - always: fsync after every write (safest, slowest)
# - everysec: fsync every second (good balance, default)
# - no: let OS decide when to sync (fastest, riskiest)
appendfsync everysec
 
# Don't fsync during BGSAVE/BGREWRITEAOF (better perf, may lose data on crash)
no-appendfsync-on-rewrite no
 
# Auto-rewrite AOF when it grows by 100% and is > 64MB
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
 
# Load truncated AOF on startup (useful after crash)
aof-load-truncated yes
 
# Use RDB-format for AOF (hybrid approach, best of both)
aof-use-rdb-preamble yes
 
# ========================================
# RECOMMENDATIONS BY USE CASE
# ========================================
 
# Pure Cache (no persistence needed):
#   save ""
#   appendonly no
 
# Standard Durability:
#   save 900 1
#   save 300 10
#   appendonly yes
#   appendfsync everysec
 
# Maximum Durability:
#   save 60 1
#   appendonly yes
#   appendfsync always
#   (Note: ~10x performance impact)
 
# Hybrid (belt and suspenders):
#   save 900 1
#   save 300 10
#   appendonly yes
#   appendfsync everysec
#   aof-use-rdb-preamble yes

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

# ========================================
# REPLICA CONFIGURATION
# ========================================
 
# On the replica server, configure to replicate from primary:
 
# Option 1: In redis.conf
replicaof 192.168.1.100 6379  # Primary host and port
 
# If primary has password
masterauth your_primary_password
 
# Serve stale data during sync (or refuse with "no")
replica-serve-stale-data yes
 
# Replica is read-only (highly recommended)
replica-read-only yes
 
# ========================================
# RUNTIME REPLICATION COMMANDS
# ========================================
 
# Make this instance a replica of another
# redis-cli
> REPLICAOF 192.168.1.100 6379
 
# Check replication status
> INFO replication
# role:slave
# master_host:192.168.1.100
# master_port:6379
# master_link_status:up
# master_last_io_seconds_ago:1
# master_sync_in_progress:0
 
# Promote replica to primary (stops replication)
> REPLICAOF NO ONE
 
# ========================================
# REDIS SENTINEL CONFIGURATION
# ========================================
 
# Sentinel monitors primary and initiates failover
# Configure in sentinel.conf:
 
# Monitor primary with name "mymaster", 
# quorum of 2 sentinels needed to confirm failover
sentinel monitor mymaster 192.168.1.100 6379 2
 
# How long to wait before considering primary down
sentinel down-after-milliseconds mymaster 30000
 
# Max replicas to reconfigure in parallel during failover
sentinel parallel-syncs mymaster 1
 
# Failover timeout
sentinel failover-timeout mymaster 180000
 
# Auth for monitored instances
sentinel auth-pass mymaster your_password

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81

# ========================================
# MEMORY MANAGEMENT
# ========================================
 
# Maximum memory Redis can use (adjust to your server)
maxmemory 4gb
 
# Eviction policy when maxmemory reached
# - noeviction: return errors for writes
# - allkeys-lru: evict least recently used keys
# - volatile-lru: evict LRU keys with TTL only
# - allkeys-lfu: evict least frequently used (Redis 4.0+)
maxmemory-policy allkeys-lru
 
# Number of keys to sample for LRU eviction
maxmemory-samples 10
 
# ========================================
# SECURITY
# ========================================
 
# Require password for all clients
requirepass your_strong_password_here
 
# Bind to specific interface (not 0.0.0.0!)
bind 127.0.0.1 10.0.0.1
 
# Disable protected mode if binding to non-localhost
protected-mode yes
 
# Rename dangerous commands (or disable with "")
rename-command FLUSHALL ""
rename-command FLUSHDB ""
rename-command DEBUG ""
rename-command KEYS "KEYS_DISABLED_IN_PROD"
 
# ========================================
# PERFORMANCE TUNING
# ========================================
 
# TCP backlog for high connection rates
tcp-backlog 511
 
# Client timeout (0 = disabled)
timeout 0
 
# TCP keepalive (seconds)
tcp-keepalive 300
 
# Number of databases (default 16, usually only use db 0)
databases 16
 
# ========================================
# LOGGING AND MONITORING
# ========================================
 
# Log level (debug, verbose, notice, warning)
loglevel notice
 
# Log file (empty = stdout)
logfile /var/log/redis/redis.log
 
# Slow log: log commands taking > 10ms
slowlog-log-slower-than 10000
 
# Keep last 128 slow commands
slowlog-max-len 128
 
# ========================================
# CLIENT CONNECTION LIMITS
# ========================================
 
# Max simultaneous clients
maxclients 10000
 
# ========================================
# LATENCY MONITORING
# ========================================
 
# Enable latency monitoring for spikes > 10ms
latency-monitor-threshold 10