Redis 速查表

# connect to a local server (default port 6379)
redis-cli

# connect with a specific host and port
redis-cli -h host.example.com -p 6380

# connect with authentication
redis-cli -a "yourpassword"
redis-cli -h host -p 6379 -a "yourpassword"

# connect over TLS
redis-cli --tls --cert client.crt --key client.key -h host -p 6379

# test the connection
PING          # returns PONG
PING "hello"  # returns "hello"

# Redis has 16 logical databases (0-15) by default
SELECT 0     # switch to database 0
SELECT 15    # switch to database 15

# move a key to another database
MOVE mykey 1

# flush the current database (DESTRUCTIVE)
FLUSHDB

# flush ALL databases (VERY DESTRUCTIVE)
FLUSHALL

# get the number of keys in the current db
DBSIZE

# get basic server info
INFO server
INFO memory
INFO replication

# set and get a key
SET user:1 "Alice"
GET user:1

# check if a key exists
EXISTS user:1          # 1 if exists, 0 if not

# delete keys
DEL user:1
DEL key1 key2 key3     # returns number deleted

# set a key with expiration (seconds)
SET session:abc "data" EX 3600

# find keys matching a pattern (avoid in production)
KEYS user:*
SCAN 0 MATCH user:* COUNT 100   # safer iteration

# rename a key
RENAME oldkey newkey

# get the type of a value stored at a key
TYPE user:1

# set expiration on an existing key
EXPIRE user:1 60        # 60 seconds
EXPIREAT user:1 1700000000  # Unix timestamp

# set expiration in milliseconds
PEXPIRE user:1 60000

# view remaining time to live (seconds)
TTL user:1              # -2 if no key, -1 if no expiry

# view TTL in milliseconds
PTTL user:1

# remove expiration (make key persistent)
PERSIST user:1

# set value AND expiration atomically
SET token "abc" EX 3600

# SCAN: cursor-based iteration (non-blocking)
SCAN 0 MATCH user:* COUNT 100
# returns: [nextCursor, [key1, key2, ...]]
SCAN <nextCursor> MATCH user:* COUNT 100
# continue until cursor returns to 0

# scan a specific type
SCAN 0 TYPE hash

# HSCAN: iterate hash fields
HSCAN myhash 0 MATCH field* COUNT 100

# SSCAN: iterate set members
SSCAN myset 0 MATCH member* COUNT 100

# ZSCAN: iterate sorted set members
ZSCAN myzset 0 MATCH member* COUNT 100

# Redis is a data structure server, not just a key-value store
# Each key holds a typed data structure:

# String: binary-safe blob (up to 512MB)
SET key "value"
SET counter 100

# List: ordered, linked-list of strings
LPUSH mylist "a" "b"
RPUSH mylist "c"

# Hash: field-value map (like a small object)
HSET user:1 name "Alice" age 30

# Set: unordered collection of unique strings
SADD tags "redis" "db"

# Sorted Set (ZSet): set scored by a float
ZADD leaderboard 100 "alice" 200 "bob"

# Stream: append-only log with IDs
XADD mystream * field value

# Bitmap, HyperLogLog, Geo: built on strings/zsets

# Use case -> data type mapping:
# Cache a string  -> String
SET api:response "json_blob"

# Store an object -> Hash (one key, multiple fields)
HSET user:1 name "Alice" email "[email protected]" age 30

# Queue / stack   -> List
LPUSH tasks "job1"   # push to head
RPOP tasks           # pop from tail (FIFO queue)

# Unique tags      -> Set
SADD post:1:tags "redis" "db" "cache"

# Leaderboard      -> Sorted Set
ZADD scores 100 "alice" 250 "bob"
ZREVRANGE scores 0 9     # top 10

# Event log        -> Stream
XADD events * type click user 1

# Count unique     -> HyperLogLog (approximate, fixed memory)
PFADD visitors "user1" "user2"

# Redis optimizes storage based on size
# Inspect the internal encoding of a key
OBJECT ENCODING mykey

# String encodings:
#   "embstr"  - short strings (< 44 bytes)
#   "raw"     - long strings
#   "int"     - integer values

# Hash/List/Set/ZSet encodings:
#   "listpack" / "intset" - small (compact, memory-efficient)
#   "hashtable" / "list" / "skiplist" - large (faster, more memory)

# check memory usage of a key
MEMORY USAGE mykey

# sample memory usage stats
MEMORY STATS

# the encoding changes automatically as the structure grows

# OBJECT subcommands for introspection
OBJECT ENCODING mykey     # internal encoding
OBJECT REFCOUNT mykey     # reference count
OBJECT IDLETIME mykey     # seconds since last access
OBJECT FREQ mykey         # access frequency (LFU mode)

# check memory usage (bytes) of a key
MEMORY USAGE mykey
MEMORY USAGE mykey SAMPLES 0   # exact count

# total memory used by the server
INFO memory | grep used_memory_human

# help understand a command
COMMAND INFO GET
COMMAND DOCS SET

# COMMAND lists all commands (huge output)
COMMAND

# each key has ONE type; using the wrong command fails
SET mykey "hello"
LPUSH mykey "x"        # WRONGTYPE error

# WRONGTYPE Operation against a key holding the wrong kind of value

# check the type before operating
TYPE mykey             # "string"
if [ "$(TYPE mykey)" = "string" ]; then ...

# TYPE returns: string|list|hash|set|zset|stream|none

# rename preserves the type
RENAME mykey newkey
TYPE newkey            # still "string"

# type is immutable; delete + recreate to change types
DEL mykey
LPUSH mykey "x"        # now a list

# set and get
SET mykey "hello"
GET mykey               # "hello"

# set only if not exists (NX) or only if exists (XX)
SET mykey "new" NX      # set only if key does NOT exist
SET mykey "new" XX      # set only if key DOES exist
SET mykey "new" EX 60 NX  # with expiry, only if new

# get and set atomically (returns old value)
GETSET counter 0        # returns old value, sets new

# append to a string
APPEND mykey " world"   # "hello world"

# get substring
GETRANGE mykey 0 4      # "hello"
SETRANGE mykey 6 "redis"# "hello redis"

# get string length
STRLEN mykey            # 11

# increment a numeric string atomically
INCR counter            # counter = counter + 1
INCRBY counter 10       # counter = counter + 10

# decrement
DECR counter            # counter = counter - 1
DECRBY counter 5        # counter = counter - 5

# floating point operations
SET price "10.50"
INCRBYFLOAT price 0.25  # "10.75"
INCRBYFLOAT price -1.00 # "9.75"

# all increment ops are atomic and concurrency-safe
# use for: counters, rate limiting, sequence IDs, stats

# set multiple keys at once (atomic)
MSET k1 "v1" k2 "v2" k3 "v3"

# get multiple keys at once
MGET k1 k2 k3           # ["v1", "v2", "v3"]

# set multiple only if ALL keys are new
MSETNX k1 "v1" k4 "v4"  # 1 if all set, 0 if any existed

# MSET/MGET reduce round-trips — much faster than
# separate SET/GET calls in a loop

# get the length of a string value
STRLEN k1               # 2

# atomic get-and-delete (useful for queues)
GETDEL mykey            # returns value AND deletes key

# bitmaps are strings operated on at the bit level
# great for boolean flags on a large user base

# set bit at position (0 or 1)
SETBIT user:1:active 7 1    # user 1 active
GETBIT user:1:active 7      # 1

# count set bits
BITCOUNT user:1:active      # total bits set to 1

# bitwise operations between strings
SETBIT users:daily:2025-01-01 5 1
SETBIT users:daily:2025-01-02 5 1
BITOP AND active-both users:daily:2025-01-01 users:daily:2025-01-02
BITCOUNT active-both        # users active on both days

# find the first set bit
BITPOS user:1:active 1      # position of first 1-bit

# BITFIELD: multiple counters in one string (compact)
# store up to 2^63 counters, each 1-64 bits, in a single key

# set a 8-bit unsigned counter at offset 0
BITFIELD mycount SET u8 0 100

# increment a counter (with overflow control)
BITFIELD mycount INCRBY u8 0 10   # returns 110
BITFIELD mycount INCRBY u8 0 10   # returns 120

# overflow control: WRAP (default), SAT, FAIL
BITFIELD mycount OVERFLOW SAT INCRBY u8 0 200  # saturates at 255
BITFIELD mycount OVERFLOW FAIL INCRBY u8 0 999 # returns nil (overflow)

# read multiple counters in one call
BITFIELD mycount GET u8 0 GET u8 8 GET u8 16

# a hash maps fields to values (like a small object)
HSET user:1 name "Alice" age 30 email "[email protected]"

# get a single field
HGET user:1 name            # "Alice"

# get multiple fields
HMGET user:1 name age       # ["Alice", "30"]

# get all fields and values
HGETALL user:1
# name  -> Alice
# age   -> 30
# email -> [email protected]

# get all field names or values
HKEYS user:1                # [name, age, email]
HVALS user:1                # [Alice, 30, [email protected]]

# get the number of fields
HLEN user:1                 # 3

# set a field only if it doesn't exist
HSETNX user:1 status "new"   # 1 if set, 0 if existed

# delete a field
HDEL user:1 email            # removes the email field

# check if a field exists
HEXISTS user:1 name          # 1 if exists, 0 if not

# increment a numeric field atomically
HINCRBY user:1 age 1         # age = 31
HINCRBY user:1 age -5        # age = 26
HINCRBYFLOAT user:1 score 0.5

# get the string length of a field's value
HSTRLEN user:1 name          # 5 (length of "Alice")

# set multiple fields (same as HSET with multiple)
HMSET user:1 a 1 b 2         # deprecated, use HSET

# HSCAN: cursor-based iteration (for large hashes)
HSCAN user:1 0 MATCH "na*" COUNT 10
# returns [cursor, [field, value, field, value, ...]]

# iterate all fields
cursor=0
while true:
  cursor, fields = HSCAN user:1 cursor COUNT 100
  process(fields)
  if cursor == 0: break

# HGETALL is fine for small hashes (< 100 fields)
# HSCAN is required for large hashes to avoid blocking

# get a random field
HRANDFIELD user:1            # one random field name
HRANDFIELD user:1 3          # 3 random field names
HRANDFIELD user:1 3 WITHVALUES  # fields + values

# before 7.4, expiration was only at the key level
EXPIRE user:1 3600    # the whole hash expires

# Redis 7.4+: field-level TTL (HEXPIRE)
HEXPIRE user:1 60 FIELDS 1 session_token
# expires only the session_token field in 60 seconds

# get TTL of a specific field
HPTTL user:1 FIELDS 1 session_token

# persist a specific field (remove its TTL)
HPERSIST user:1 FIELDS 1 session_token

# get all fields and their expiration times
HEXPIRETIME user:1 FIELDS 1 session_token

# 1. Store an object (user profile, config)
HSET config:app name "MyApp" version "2.0" debug "off"

# 2. Per-user counters
HINCRBY user:1:stats logins 1
HINCRBY user:1:stats page_views 1
HGET user:1:stats logins

# 3. Shopping cart (product_id -> quantity)
HSET cart:user:1 product:100 2
HINCRBY cart:user:1 product:100 1
HDEL cart:user:1 product:100

# 4. Feature flags
HSET feature_flags user:1 new_ui 1 beta 0
HGET feature_flags user:1 new_ui   # 1

# 5. Group small key-value pairs into one hash (memory savings)
# instead of SET k1 v1; SET k2 v2 -> HSET bucket k1 v1 k2 v2

# lists are ordered sequences of strings (linked lists)
# push to head (left) or tail (right)
LPUSH mylist "a" "b"      # list: [b, a]
RPUSH mylist "c"          # list: [b, a, c]

# pop from head or tail
LPOP mylist               # "b" (removes from head)
RPOP mylist               # "c" (removes from tail)

# pop multiple
LPOP mylist 2             # ["a", ...]

# get length
LLEN mylist               # 0

# blocking pop (waits if empty, up to timeout)
BLPOP queue:tasks 30      # blocks up to 30 seconds
BRPOP queue:tasks 0       # blocks forever

# get element by index (0-based, O(N) traversal)
LINDEX mylist 0           # first element
LINDEX mylist -1          # last element

# get a range of elements
LRANGE mylist 0 -1        # all elements
LRANGE mylist 0 2         # first 3 elements
LRANGE mylist -3 -1       # last 3 elements

# set an element by index
LSET mylist 1 "new"

# get length
LLEN mylist

# trim to a range (removes everything outside)
LTRIM mylist 0 99         # keep only first 100 elements

# remove elements by value
LREM mylist 2 "value"     # remove first 2 occurrences of "value"
LREM mylist -2 "value"    # remove last 2 occurrences

# simple FIFO queue: LPUSH to add, RPOP to consume
LPUSH tasks "job1" "job2"
RPOP tasks                # "job1" (FIFO order)

# reliable queue with BRPOP (blocking, waits for work)
BRPOP tasks 0             # blocks until a task is available

# move items between lists atomically
LPUSH source "item"
RPOPLPUSH source destination  # move one item

# blocking move (reliable queue pattern)
BRPOPLPUSH source destination 30  # blocks up to 30s

# Redis 6.2+: BLMOVE replaces BRPOPLPUSH
BLMOVE source destination RIGHT LEFT 30

# capped queue (e.g., recent events)
LPUSH recent "event"
LTRIM recent 0 99         # keep latest 100

# insert before or after a pivot element
LINSERT mylist BEFORE "pivot" "new"
LINSERT mylist AFTER "pivot" "new"

# remove elements by value
LREM mylist 1 "value"     # remove first occurrence from head
LREM mylist -1 "value"    # remove first occurrence from tail
LREM mylist 0 "value"     # remove ALL occurrences

# pop and push in one atomic operation
RPOPLPUSH source dest

# get and remove from both ends (6.2+)
LPOP mylist 2             # pop 2 from head
LMPOP 2 mylist LEFT COUNT 3  # (7.0+) pop from one of multiple lists

# find the position of an element (6.0+)
LPOS mylist "value"
LPOS mylist "value" RANK 2   # find 2nd occurrence

# 1. Task queue (FIFO)
LPUSH tasks "process_order:123"
BRPOP tasks 30

# 2. Recent activity feed (capped)
LPUSH user:1:feed "post:99" "post:98"
LTRIM user:1:feed 0 49     # latest 50

# 3. Stack (LIFO)
LPUSH stack "item"
LPOP stack

# 4. Circular buffer (limited size)
LPUSH buffer "new"
LTRIM buffer 0 999         # keep 1000 items

# 5. Rate limiter (sliding window)
LPUSH rate:user:1 <timestamp>
LTRIM rate:user:1 0 99
LLEN rate:user:1           # request count in window

# sets: unordered collection of unique strings
SADD tags "redis" "db" "cache"
SADD tags "redis"          # ignored (already exists)

# remove a member
SREM tags "cache"

# check membership
SISMEMBER tags "redis"     # 1 if member, 0 if not

# get the number of members
SCARD tags                 # 2

# get all members (avoid on large sets)
SMEMBERS tags              # ["redis", "db"]

# get a random member
SRANDMEMBER tags           # one random member
SRANDMEMBER tags 3         # 3 random (may repeat)
SRANDMEMBER tags -3        # 3 unique random

# pop a random member (removes it)
SPOP tags                  # removes and returns one

# set union, intersection, difference
SADD set1 "a" "b" "c"
SADD set2 "b" "c" "d"

# union (all unique members from all sets)
SUNION set1 set2           # a, b, c, d

# intersection (members in ALL sets)
SINTER set1 set2           # b, c

# difference (members in set1 but NOT in set2)
SDIFF set1 set2            # a

# store the result in a new set
SUNIONSTORE result set1 set2
SINTERSTORE result set1 set2
SDIFFSTORE result set1 set2

# count intersection without returning members (7.0+)
SINTERCARD 2 set1 set2 LIMIT 0

# SSCAN: cursor-based iteration for large sets
SSCAN myset 0 MATCH "pre*" COUNT 100
# returns [cursor, [member1, member2, ...]]

# move a member from one set to another atomically
SMOVE source dest "member"

# pop multiple random members (6.2+)
SPOP myset 3               # removes and returns 3 members

# blocking set pop (7.0+) - waits until a member exists
BSPOP myset 30             # blocks up to 30 seconds

# check multiple memberships at once (7.4+)
SMISMEMBER myset "a" "b" "c"
# returns [1, 0, 1]

# 1. Tags / categories
SADD post:1:tags "redis" "db"
SADD post:2:tags "redis" "cache"
SINTER post:1:tags post:2:tags   # posts sharing "redis"

# 2. Unique visitors per day
SADD visitors:2025-01-01 "user:1" "user:2"
SCARD visitors:2025-01-01        # visitor count

# 3. Followers / following
SADD user:1:following "user:2" "user:3"
SADD user:2:following "user:1" "user:3"
SINTER user:1:following user:2:following  # mutual follows

# 4. Blacklist / whitelist
SADD blacklist "ip:1.2.3.4"
SISMEMBER blacklist "ip:1.2.3.4"

# 5. Lottery / random selection
SADD participants "u1" "u2" "u3" "u4"
SPOP participants 1        # draw a winner

# Set: unordered, unique, O(1) membership check
SADD myset "a" "b"
SISMEMBER myset "a"       # O(1)

# Sorted Set: unique + ordered by score
ZADD myzset 1 "a" 2 "b"
ZRANK myzset "a"          # 0 (rank)

# List: ordered, allows duplicates, O(N) search
LPUSH mylist "a" "a"
LPOS mylist "a"           # 0 (first match)

# Hash: field-value pairs, no ordering
HSET myhash a 1 b 2

# Choose:
# - need uniqueness only?           -> Set
# - need uniqueness + ordering?     -> Sorted Set
# - need ordering + allow dups?     -> List
# - need key-value mapping?         -> Hash

# sorted set: unique members ordered by a float score
ZADD leaderboard 100 "alice" 200 "bob" 150 "carol"

# update a score (re-add with new score)
ZADD leaderboard 250 "alice"   # alice's score becomes 250

# add with options (NX: only new, XX: only existing, GT/LT: conditional)
ZADD leaderboard NX 300 "new"
ZADD leaderboard GT 300 "alice"  # only if new score is greater

# get a member's score
ZSCORE leaderboard "alice"      # "250"

# get a member's rank (0-based, ascending)
ZRANK leaderboard "alice"       # rank from lowest
ZREVRANK leaderboard "alice"    # rank from highest (0 = top)

# get the number of members
ZCARD leaderboard               # 3

# get members by index range (ascending)
ZRANGE leaderboard 0 -1              # all, ascending
ZRANGE leaderboard 0 2               # first 3 (lowest scores)
ZRANGE leaderboard -3 -1             # last 3 (highest scores)

# get members by score range
ZRANGEBYSCORE leaderboard 100 200    # scores 100-200
ZRANGEBYSCORE leaderboard 100 +inf   # scores >= 100
ZRANGEBYSCORE leaderboard (100 200   # scores > 100 (exclusive)

# descending order
ZREVRANGE leaderboard 0 2            # top 3 (highest scores)
ZREVRANGEBYSCORE leaderboard 200 100 # 200 down to 100

# with scores
ZRANGE leaderboard 0 -1 WITHSCORES

# Redis 6.2+: unified ZRANGE with BYSCORE/REV options
ZRANGE leaderboard 100 200 BYSCORE
ZRANGE leaderboard 200 100 BYSCORE REV

# increment a member's score
ZINCRBY leaderboard 50 "alice"     # alice += 50

# get the score
ZSCORE leaderboard "alice"

# remove a member
ZREM leaderboard "bob"

# remove members by rank range
ZREMRANGEBYRANK leaderboard 0 9    # remove lowest 10
ZREMRANGEBYSCORE leaderboard 0 100 # remove scores <= 100

# count members in a score range
ZCOUNT leaderboard 100 200

# get the rank
ZRANK leaderboard "alice"          # ascending rank
ZREVRANK leaderboard "alice"       # descending rank (0 = highest)

# get multiple members' scores at once
ZMSCORE leaderboard "alice" "bob"  # ["250", "200"]

# when all scores are equal, sorted sets act as a sorted string set
ZADD myset 0 "apple" 0 "banana" 0 "cherry" 0 "date"

# range by lexicographical order
ZRANGEBYLEX myset "[a" "[c"        # apple, banana, cherry
ZRANGEBYLEX myset "(a" "[c"        # banana, cherry (a exclusive)
ZRANGEBYLEX myset "-" "[b"         # from start to banana
ZRANGEBYLEX myset "[c" "+"         # cherry to end

# count in a lex range
ZLEXCOUNT myset "[a" "[c"          # 3

# remove by lex range
ZREMRANGEBYLEX myset "[a" "[b"

# Redis 6.2+: use ZRANGE with BYLEX
ZRANGE myset "[a" "[c" BYLEX

# ZUNIONSTORE / ZINTERSTORE: combine multiple sorted sets
ZADD set1 1 "a" 2 "b"
ZADD set2 2 "a" 3 "c"

# union: sum scores by default (a = 1+2 = 3)
ZUNIONSTORE result 2 set1 set2
ZRANGE result 0 -1 WITHSCORES   # a=3, b=2, c=3

# specify aggregation: SUM (default), MIN, MAX
ZUNIONSTORE result 2 set1 set2 AGGREGATE MAX   # a=2, b=2, c=3

# apply weights to each set's scores
ZUNIONSTORE result 2 set1 set2 WEIGHTS 2 1     # set1 scores * 2

# intersection (only members in ALL sets)
ZINTERSTORE result 2 set1 set2                 # only "a"

# Redis 6.2+: return without storing
ZUNION 2 set1 set2 WITHSCORES
ZINTER 2 set1 set2 WITHSCORES
ZDIFF 2 set1 set2 WITHSCORES

# subscribe to channels (in one client)
SUBSCRIBE news alerts

# publish a message to a channel (from another client)
PUBLISH news "Breaking: Redis 7.4 released"

# pattern subscription (glob-style)
PSUBSCRIBE news.*

# unsubscribe
UNSUBSCRIBE news
PUNSUBSCRIBE news.*

# note: published messages are NOT persisted
# if no subscriber is listening, the message is lost
# for persistent messaging, use Streams instead

# regular Pub/Sub broadcasts to ALL cluster nodes (expensive)
# sharded Pub/Sub routes messages only to the shard owning the channel

# subscribe to a shard channel
SSUBSCRIBE mychannel

# publish to a shard channel
SPUBLISH mychannel "hello"

# sharded pub/sub uses the channel name for shard routing
# so messages only travel within one shard (more efficient)

# use sharded pub/sub in a Redis Cluster to avoid
# broadcasting messages to every node

# regular SUBSCRIBE/PUBLISH still work in cluster mode
# but they broadcast to all shards (cluster-wide overhead)

# 1. Real-time chat
# client subscribes to their room
SUBSCRIBE chat:room:42
# server publishes messages
PUBLISH chat:room:42 "Alice: hello"

# 2. Cache invalidation
# subscribe to invalidation channel
SUBSCRIBE cache:invalidate
# when data changes, publish
PUBLISH cache:invalidate "user:123"

# 3. Event notifications (with PSUBSCRIBE patterns)
PSUBSCRIBE user:*:login
# matches user:1:login, user:2:login, etc.

# 4. Live dashboards
SUBSCRIBE metrics:*
# push real-time updates to dashboards

# 5. Multiplayer game state
PUBLISH game:room:1 "player_moved:5:10"

# enable keyspace notifications (disabled by default)
CONFIG SET notify-keyspace-events "KEA"
# K = keyspace events, E = keyevent events
# g = generic (DEL, EXPIRE, ...), $ = string, l = list
# h = hash, s = set, z = sorted set, x = expired, e = evicted
# A = all except m (m = keymiss)

# subscribe to events on a specific key
SUBSCRIBE __keyspace@0__:mykey
# publishes "set", "del", "expire", etc.

# subscribe to events of a specific type
SUBSCRIBE __keyevent@0__:del
# publishes the key name that was deleted

# example: react when a key expires
SUBSCRIBE __keyspace@0__:session:abc
# receives "expired" when the key TTL ends

# Pub/Sub: fire-and-forget, no persistence
PUBLISH channel "msg"   # lost if no subscribers

# Streams: persistent, replayable, consumer groups
XADD mystream * msg "hello"   # stored forever (or capped)

# Pub/Sub pros: simple, low latency, fan-out
# Pub/Sub cons: no persistence, no replay, no consumer groups

# Streams pros: persistent, replayable, consumer groups, ACK
# Streams cons: more complex, requires cleanup (XADD MAXLEN)

# when to use each:
# - real-time notifications (loss OK)     -> Pub/Sub
# - chat with presence                    -> Pub/Sub + Streams
# - reliable task queue                   -> Streams
# - event sourcing / audit log            -> Streams
# - fan-out to transient subscribers      -> Pub/Sub

# MULTI starts a transaction; commands are queued
MULTI
SET counter 1
INCR counter
INCR counter
GET counter
EXEC
# returns: [OK, 2, 3, "3"]

# discard a transaction (cancels all queued commands)
MULTI
SET x 1
DISCARD

# errors during queuing (syntax) abort the whole transaction
# errors during EXEC (e.g., wrong type) skip only that command

# WATCH monitors keys; if any change before EXEC, the transaction aborts
WATCH counter
val = GET counter          # read current value
MULTI
INCR counter               # queue the update
EXEC                       # nil if counter changed -> retry

# typical pattern: read-modify-write with retry
WATCH key
current = GET key
new_value = compute(current)
MULTI
SET key new_value
EXEC
# if EXEC returns nil, the key changed — loop and retry

# UNWATCH cancels all watches
UNWATCH

# pipelining sends multiple commands in one network round-trip
# (no atomicity guarantee, just reduced latency)

# in a client library (e.g., Node.js / redis):
const pipe = client.multi()
pipe.set("k1", "v1")
pipe.set("k2", "v2")
pipe.incr("counter")
const results = await pipe.exec()

# vs MULTI/EXEC: pipelining does NOT guarantee atomicity
# other clients can interleave between pipelined commands

# raw protocol example (RESP):
# *1\r\n$4\r\nPING\r\n*1\r\n$4\r\nPING\r\n
# (send multiple commands at once, read multiple replies)

# Lua scripts run atomically — no other command runs during execution
# EVAL: run a script
EVAL "return redis.call('GET', KEYS[1])" 1 mykey

# a check-and-set script:
EVAL "
  local cur = redis.call('GET', KEYS[1])
  if cur == ARGV[1] then
    return redis.call('SET', KEYS[1], ARGV[2])
  end
  return 0
" 1 mykey "expected" "newvalue"

# load a script once, then run by SHA hash (efficient)
SCRIPT LOAD "return redis.call('GET', KEYS[1])"
# returns a SHA1 hash
EVALSHA <sha1> 1 mykey

# check if a script is loaded
SCRIPT EXISTS <sha1>

# FUNCTIONS are named, versioned Lua scripts (better than EVAL)
# register a function library
FUNCTION LOAD '#!lua name=mylib
redis.register_function("my_set",
  function(keys, args)
    return redis.call("SET", keys[1], args[1])
  end
)'

# call a function by name
FCALL my_set 1 mykey "hello"

# list loaded libraries
FUNCTION LIST

# get the function's code
FUNCTION DUMP mylib

# delete a library
FUNCTION DELETE mylib

# functions persist across restarts (unlike EVAL scripts)

# RDB: point-in-time snapshot of the dataset (compact binary)
# trigger a snapshot manually:
SAVE              # blocks until done (use with care)
BGSAVE            # forks a background process (non-blocking)
LASTSAVE          # Unix timestamp of the last successful save

# configure automatic snapshots in redis.conf:
#   save 3600 1     # save if >= 1 key changed in 3600s
#   save 300 100    # save if >= 100 keys changed in 300s
#   save 60 10000   # save if >= 10000 keys changed in 60s
#   save ""         # disable RDB entirely

# RDB file location (default: dump.rdb in the working dir)
CONFIG GET dbfilename
CONFIG GET dir

# RDB pros: compact, fast restart, great for backups
# RDB cons: potential data loss between snapshots

# AOF: logs every write command (durable, replayable)
# enable in redis.conf:
#   appendonly yes
#   appendfilename "appendonly.aof"

# fsync policies:
#   appendfsync always    # fsync every write (safest, slowest)
#   appendfsync everysec  # fsync once per second (default, balanced)
#   appendfsync no        # let OS decide (fastest, risk of data loss)

# trigger AOF rewrite (compacts the log)
BGREWRITEAOF

# AOF auto-rewrite thresholds:
#   auto-aof-rewrite-percentage 100
#   auto-aof-rewrite-min-size 64mb

# AOF pros: minimal data loss (1 sec max with everysec)
# AOF cons: larger files, slower load on restart

# use BOTH RDB and AOF for the best of both worlds
# redis.conf:
#   save 900 1          # RDB snapshots
#   appendonly yes      # AOF enabled
#   aof-use-rdb-preamble yes  # RDB header in AOF (faster load)

# on restart, Redis loads AOF (more complete)
# the AOF file can start with an RDB snapshot (faster load)
# followed by incremental AOF commands

# disaster recovery: keep both RDB and AOF backups
# RDB for fast restores, AOF for minimal data loss

# check what's enabled
CONFIG GET save
CONFIG GET appendonly

# monitor persistence operations
INFO persistence

# backup: copy the RDB file while Redis runs
cp /var/lib/redis/dump.rdb /backup/dump-$(date +%F).rdb

# or trigger a BGSAVE first for a consistent snapshot
BGSAVE
# wait for LASTSAVE to update, then copy the file

# AOF backup: copy the AOF manifest + files
cp /var/lib/redis/appendonly.aof.* /backup/

# restore: stop Redis, replace the data files, start
redis-cli SHUTDOWN
cp /backup/dump.rdb /var/lib/redis/dump.rdb
redis-server /etc/redis/redis.conf

# for point-in-time recovery from AOF:
# edit the AOF file to remove unwanted commands, then restart

# RDB file analysis (offline)
redis-check-rdb /var/lib/redis/dump.rdb
redis-check-aof /var/lib/redis/appendonly.aof

# choosing a persistence strategy:

# 1. RDB only (cache, re-buildable data)
#    save 3600 1
#    appendonly no
#    -> fast restart, some data loss on crash

# 2. AOF only (durability-critical, everysec)
#    save ""
#    appendonly yes
#    appendfsync everysec
#    -> up to 1s data loss, larger files

# 3. RDB + AOF (recommended production default)
#    save 3600 1
#    appendonly yes
#    aof-use-rdb-preamble yes
#    -> durability + fast restart

# 4. No persistence (pure cache, ephemeral)
#    save ""
#    appendonly no
#    -> maximum speed, all data lost on restart

# 5. Replica-only persistence (offload from primary)
#    primary: no persistence (max performance)
#    replica: RDB + AOF (persist here instead)

# Redis Cluster shards data across multiple nodes (16,384 slots)
# each key maps to a slot: CRC16(key) % 16384

# create a cluster (6 nodes: 3 primaries + 3 replicas)
redis-cli --cluster create \
  host1:7000 host2:7000 host3:7000 \
  host1:7001 host2:7001 host3:7001 \
  --cluster-replicas 1

# check cluster status
redis-cli -p 7000 CLUSTER INFO
redis-cli -p 7000 CLUSTER NODES

# count slots assigned
CLUSTER COUNTKEYSINSLOT 1234

# get the slot for a key
CLUSTER KEYSLOT mykey   # e.g., 1234

# multi-key operations require keys in the SAME slot
# use hash tags to force keys to the same slot:
SET {user:1}:profile "alice"
SET {user:1}:cart "items"
# both map to the same slot (the part inside {})

# hash tags: the substring inside {} determines the slot
# {user:1}:profile and {user:1}:cart share a slot
# enabling multi-key operations:

MSET {user:1}:name "Alice" {user:1}:age 30   # OK (same slot)
MSET user:1:name "Alice" user:2:age 30       # CROSSSLOT error

# transactions across keys (must be same slot)
WATCH {order:1}:items
MULTI
HINCRBY {order:1}:items product:1 1
HINCRBY {order:1}:total 10
EXEC

# Lua scripts accessing multiple keys (all keys must share a slot)
EVAL "..." 2 {user:1}:a {user:1}:b

# design hash tags carefully:
# - too broad -> hotspots (one shard overloaded)
# - too narrow -> can't do multi-key ops

# move slots from one node to another (resharding)
redis-cli --cluster reshard host1:7000
# prompts for: how many slots, target node, source nodes

# add a new node to the cluster
redis-cli --cluster add-node newhost:7000 host1:7000

# add a new replica to an existing primary
redis-cli --cluster add-node newhost:7001 host1:7000 \
  --cluster-slave --cluster-master-id <nodeId>

# remove a node (migrates its slots first)
redis-cli --cluster del-node host1:7000 <nodeId>

# rebalance slots across all nodes
redis-cli --cluster rebalance host1:7000

# check cluster health
redis-cli --cluster check host1:7000
redis-cli --cluster fix host1:7000   # repair issues

# when a primary fails, its replica is promoted automatically
# manual failover (for maintenance):
redis-cli -p 7001 CLUSTER FAILOVER         # replica -> primary
redis-cli -p 7001 CLUSTER FAILOVER FORCE   # without primary agreement
redis-cli -p 7001 CLUSTER FAILOVER TAKEOVER # force, may split brain

# view the cluster topology
CLUSTER NODES
CLUSTER SHARDS    # (7.0+)

# a node flagged as FAIL after timeout is removed
# cluster- node-timeout 15000   (15 seconds in redis.conf)

# if a cluster has no replica for a failing primary,
# that slot range becomes unavailable until the primary returns
# (or you manually fix it)

# CLUSTER RESET resets a node (HARD removes all data)

# Single Redis: one instance, no HA, no scaling
#    simplest, fine for dev/small cache

# Sentinel: 1 primary + replicas + sentinel monitors
#    HA (automatic failover), no sharding
#    good when dataset fits on one machine
#    sentinel.conf: sentinel monitor mymaster host 6379 2

# Cluster: sharded across N primaries + replicas
#    HA + horizontal scaling
#    good for large datasets / high throughput
#    more complex (multi-key limitations)

# choose:
# - dev / tiny cache              -> single
# - production, data fits in RAM  -> sentinel (HA)
# - data > single machine RAM     -> cluster (sharding)
# - very high write throughput    -> cluster (shard writes)

# managed Redis (Redis Cloud, ElastiCache) handles
# sentinel/cluster ops for you

# Sentinel monitors a primary + replicas, handles failover
# sentinel.conf (port 26379):
sentinel monitor mymaster 192.168.1.10 6379 2
# name, host, port, quorum (number of sentinels to agree on failure)

sentinel down-after-milliseconds mymaster 30000   # 30s to declare down
sentinel failover-timeout mymaster 180000         # 3min failover timeout
sentinel parallel-syncs mymaster 1                # replicas to resync in parallel

# start a sentinel:
redis-sentinel /etc/redis/sentinel.conf
# or: redis-server /etc/redis/sentinel.conf --sentinel

# deploy at least 3 sentinels (odd number) for quorum
# sentinels communicate with each other and the Redis nodes

# connect to a sentinel to find the current primary:
redis-cli -p 26379 SENTINEL get-master-addr-by-name mymaster

# query sentinel state
redis-cli -p 26379 SENTINEL masters          # all monitored primaries
redis-cli -p 26379 SENTINEL master mymaster  # details of one primary
redis-cli -p 26379 SENTINEL replicas mymaster
redis-cli -p 26379 SENTINEL sentinels mymaster

# find the current primary address
SENTINEL get-master-addr-by-name mymaster

# force a failover (manual, for maintenance)
SENTINEL failover mymaster

# reset a master's state (clears replicas/sentinels discovered)
SENTINEL reset mymaster

# check the number of OK sentinels
SENTINEL ckquorum mymaster
# "OK 3 usable Sentinels" or "NOQUORUM" warning

# clients should connect to sentinels, NOT directly to the primary
# libraries: redis-py, jedis, ioredis all support sentinel mode

# when a primary is down longer than down-after-milliseconds:
# 1. sentinels mark it SDOWN (subjectively down)
# 2. quorum agrees -> ODOWN (objectively down)
# 3. a sentinel is elected leader
# 4. leader picks the best replica (most up-to-date)
# 5. promotes the replica: REPLICAOF NO ONE
# 6. reconfigures other replicas to follow the new primary
# 7. updates sentinel configs with the new primary address
# 8. clients reconnect via sentinel discovery

# during failover, writes fail briefly (seconds)
# reads from replicas continue (if the app uses them)

# the old primary, when it returns, is reconfigured as a replica
# of the new primary (its writes during partition are lost)

# monitor failover events:
redis-cli -p 26379 SUBSCRIBE +switch-master
redis-cli -p 26379 SUBSCRIBE +failover-*

# replicas can serve reads (eventually consistent)
REPLICAOF 192.168.1.10 6379    # become a replica of this primary
REPLICAOF NO ONE               # promote to primary (manual)

# check replication status
INFO replication
# role:slave, master_host, master_link_status:up, ...

# replica lag (bytes and seconds lagging behind primary)
# master_repl_offset vs slave_repl_offset

# clients can route reads to replicas for read scaling
# but beware: replica data may be stale (replication lag)

# Redis Sentinel + clients: configure read preference
#   READONLY command enables reads on a replica

# writable replicas (Redis 4.0+): accept writes that are
# local-only (not replicated) — useful for ephemeral data
#   replica-read-only no   (in redis.conf)

# replication is asynchronous: primary -> replica (one-way)
# initial sync: full RDB transfer + buffered commands
# subsequent: streaming command log

# trigger a full resync (force replica to reload everything)
redis-cli -p 6380 REPLICAOF NO ONE
redis-cli -p 6380 REPLICAOF 192.168.1.10 6379

# partial resync (PSYNC): if the replica disconnects briefly,
# it can resume from the replication backlog (no full resync)
#   repl-backlog-size 1mb    (tune for longer disconnects)

# monitor replication in real-time
redis-cli -p 6379 INFO replication
# master_repl_offset, slave_read_repl_offset, backlog_size

# chain replication: a replica of a replica (tree topology)
# reduces load on the primary for many replicas
REPLICAOF replica-host 6380   # chain off another replica

# replication ID changes on every primary restart -> full resync

# streams are append-only logs with IDs (durable, replayable)
# XADD: add an entry (auto-generate ID with *)
XADD mystream * sensor temp 25.5
# returns an ID like "1700000000000-0" (timestamp-sequence)

# add with a specific ID (must be greater than the last)
XADD mystream 1700000001000-0 sensor temp 26.0

# read entries
XREAD COUNT 10 STREAMS mystream 0        # from the beginning
XREAD COUNT 10 STREAMS mystream $        # new entries only (tail)

# block for new entries (like BLPOP for lists)
XREAD BLOCK 30000 STREAMS mystream $     # blocks up to 30s

# get the length
XLEN mystream                            # number of entries

# capped stream (keep only last N entries)
XADD mystream MAXLEN 1000 * field value
XADD mystream MAXLEN ~ 1000 * field value  # ~ = approx (faster)

# create a consumer group at a specific position
XGROUP CREATE mystream mygroup $        # only new entries
XGROUP CREATE mystream mygroup 0        # all existing entries
XGROUP CREATE mystream mygroup 0 MKSTREAM  # create stream if missing

# read entries as a consumer (assigns pending entries)
XREADGROUP GROUP mygroup consumer1 COUNT 10 STREAMS mystream >
# '>' = never-delivered entries

# read pending entries (for this consumer, not yet acked)
XREADGROUP GROUP mygroup consumer1 STREAMS mystream 0

# acknowledge processing (removes from pending)
XACK mystream mygroup <id1> <id2>

# view pending entries (pending entries list)
XPENDING mystream mygroup
# returns: count, min-id, max-id, consumers

# view a consumer's pending entries
XPENDING mystream mygroup - + 10 consumer1

# XCLAIM: reassign a pending entry to another consumer
# (after the original consumer died or is too slow)
XCLAIM mystream mygroup consumer2 60000 <entry-id>
# 60000 = min idle time (ms) before claiming

# XAUTOCLAIM (6.2+): auto-claim stale pending entries
XAUTOCLAIM mystream mygroup consumer2 60000 0 COUNT 10
# claims entries idle > 60s, returns them to consumer2

# view stream info
XINFO STREAM mystream FULL
XINFO GROUPS mystream
XINFO CONSUMERS mystream mygroup

# delete an entry (rarely needed)
XDEL mystream <id>

# trim the stream (free memory)
XTRIM mystream MAXLEN 10000
XTRIM mystream MINID 1700000000000   # remove older than this ID

# range queries
XRANGE mystream - +                   # all entries
XRANGE mystream 1700000000000 - 1700000009999 -

# 1. Reliable task queue (replaces lists + BRPOPLPUSH)
XADD tasks * type email payload '{"to":"[email protected]"}'
XREADGROUP GROUP workers worker1 COUNT 1 STREAMS tasks >
# process and XACK on success

# 2. Event sourcing / audit log
XADD events * user 1 action login ip 1.2.3.4
# replay with XRANGE for reconstruction

# 3. Real-time analytics (time-windowed)
XADD metrics * cpu 75 mem 60
XRANGE metrics - +           # query by time range

# 4. Chat history (capped)
XADD chat:room:1 MAXLEN ~ 1000 * user alice msg "hello"
XREAD COUNT 20 STREAMS chat:room:1 0

# 5. IoT sensor data (with TTL via trimming)
XADD sensors:temp MAXLEN ~ 86400 * value 22.5  # keep last day

# Redis Streams vs Lists (for queues):
# Lists: LPUSH/BRPOP, simple, NO retry/recovery
# Streams: consumer groups, ACK, pending, XAUTOCLAIM

# Redis Streams vs Kafka:
# Streams: in-memory (RAM-bounded), simpler, single-node or cluster
# Kafka: disk-based, partitioned, higher throughput at scale

# Redis Streams vs RabbitMQ:
# Streams: simpler, fewer features, in-memory
# RabbitMQ: rich routing, exchanges, acknowledgments, disk

# Streams capacity: limited by RAM (use MAXLEN/MINID to trim)
# Streams persistence: RDB/AOF (like any Redis data)

# choose Streams when:
# - you need a queue but already use Redis
# - dataset fits in RAM
# - you want simplicity over Kafka's scale
# - you need consumer groups with retry

# HyperLogLog: approximate unique count with fixed 12KB memory
# perfect for counting unique visitors, IPs, etc. at scale

# add elements (returns 1 if the cardinality changed)
PFADD visitors:2025-01-01 "user:1" "user:2" "user:3"

# get the approximate unique count
PFCOUNT visitors:2025-01-01        # ~3

# merge multiple HLLs (e.g., monthly uniques from daily)
PFMERGE visitors:2025-01 visitors:2025-01-01 visitors:2025-01-02

# accuracy: ~0.81% standard error
# memory: always ~12KB regardless of element count (up to 2^64)

# vs a Set for unique counting:
# SADD + SCARD: exact, but O(N) memory (grows with count)
# PFADD + PFCOUNT: approximate, fixed 12KB memory
# 1 million uniques: Set ~10MB, HLL always 12KB

# bitmaps: boolean operations on a string's bits
# ideal for per-user flags over a large user base (user ID = bit position)

# mark user 7 as "active today"
SETBIT active:2025-01-01 7 1

# check if user 7 was active
GETBIT active:2025-01-01 7          # 1

# count active users
BITCOUNT active:2025-01-01

# users active on BOTH days (AND)
BITOP AND active:both active:2025-01-01 active:2025-01-02
BITCOUNT active:both

# users active on EITHER day (OR)
BITOP OR active:either active:2025-01-01 active:2025-01-02

# users active on day 1 but NOT day 2 (XOR or AND NOT)
BITOP XOR diff active:2025-01-01 active:2025-01-02

# find the first active user
BITPOS active:2025-01-01 1         # position of first 1-bit

# memory: 1 million users = 125KB (1 bit per user)

# GEO: built on sorted sets, stores (lng, lat, name) points
# add locations (longitude, latitude, name)
GEOADD places -73.99 40.73 "Park" -73.98 40.75 "Museum" -74.00 40.70 "Cafe"

# get coordinates of a member
GEOPOS places "Park"               # [["-73.99","40.73"]]

# distance between two members (meters by default)
GEODIST places "Park" "Museum"     # "1623.5"
GEODIST places "Park" "Museum" km  # "1.623"

# find members within a radius (meters)
GEOSEARCH places FROMMEMBER "Park" BYRADIUS 1000 m ASC
# members within 1km of Park, sorted by distance

# find members within a bounding box
GEOSEARCH places FROMLONLAT -73.99 40.73 BYBOX 2 2 km

# GEOSEARCH also returns distance and coordinates
GEOSEARCH places FROMMEMBER "Park" BYRADIUS 500 m WITHCOORD WITHDIST

# add multiple points at once
GEOADD cities -122.41 37.78 "San Francisco" -74.00 40.71 "New York"

# get the geohash of a member
GEOHASH places "Park"              # ["dr5reg..."]

# remove a member (uses ZREM internally)
ZREM places "Cafe"

# count members in an area (using ZCOUNT on the underlying zset)
ZCOUNT places -180 180             # all members

# GEOSEARCH with a center point and limiting results
GEOSEARCH places FROMLONLAT -73.99 40.73 \
  BYRADIUS 2000 m ASC COUNT 5      # nearest 5 within 2km

# GEOSEARCHSTORE: store results in a new key (6.2+)
GEOSEARCHSTORE nearby FROMLONLAT -73.99 40.73 BYRADIUS 1000 m

# the underlying structure is a sorted set, so:
ZRANGE places 0 -1                 # list all members
ZSCORE places "Park"               # geohash score

# Cardinality counting (unique elements):
#   exact, need membership test  -> Set (SADD/SISMEMBER/SCARD)
#   approximate, huge scale      -> HyperLogLog (PFADD/PFCOUNT)
#   per-time-bucket uniques      -> HLL per day, PFMERGE for ranges

# Boolean flags over a dense ID space:
#   small number of flags        -> Set of active IDs
#   millions of users            -> Bitmap (SETBIT/BITCOUNT)
#   multiple counters compactly  -> BITFIELD

# Geographic points:
#   "find nearby" queries        -> GEO (GEOADD/GEOSEARCH)
#   complex GIS                  -> use PostGIS/external

# Time-series data:
#   simple, capped               -> Stream (XADD MAXLEN)
#   dedicated TS module          -> RedisTimeSeries (module)

# Probabilistic structures (modules):
#   Bloom filter                 -> RedisBloom (BF.ADD/BF.EXISTS)
#   Top-K                        -> RedisBloom (TOPK)

# set a password (simple, legacy)
CONFIG SET requirepass "strongpassword"
AUTH "strongpassword"

# Redis 6+: ACL system with users and permissions
# add a user with limited access
ACL SETUSER appuser on >secretpassword ~app:* +get +set +del -@dangerous

# list users
ACL LIST
ACL WHOAMI                # current user

# view a user's permissions
ACL GETUSER appuser

# delete a user
ACL DELUSER appuser

# save ACLs to a file (persists across restarts)
ACL SAVE
ACL LOAD                  # reload from file

# aclfile in redis.conf:
#   aclfile /etc/redis/users.acl

# enable TLS in redis.conf (Redis 6+)
# generate certs first (or use a real CA):
port 0                       # disable plain-text port
tls-port 6379
tls-cert-file /etc/redis/redis.crt
tls-key-file /etc/redis/redis.key
tls-ca-cert-file /etc/redis/ca.crt

# connect with TLS
redis-cli --tls --cert client.crt --key client.key \
  --cacert ca.crt -h host -p 6379

# mutual TLS (mTLS) for client authentication
tls-cluster yes              # TLS between cluster nodes
tls-replication yes          # TLS between primary/replica
tls-auth-clients yes         # require client certs

# mix TLS and plain text (transition period)
# port 6380
# tls-port 6379

# bind to specific interfaces (never 0.0.0.0 in prod)
bind 127.0.0.1 10.0.0.5
protected-mode yes            # blocks external access if no password

# require authentication
requirepass "strong-random-password"

# rename dangerous commands (or disable them)
rename-command FLUSHALL ""
rename-command FLUSHDB ""
rename-command CONFIG "CONFIG_9f8a7b"
rename-command KEYS ""

# Redis 6+: better to use ACLs instead of rename-command
# ACL SETUSER default -flushall -config -keys

# disable debug commands
# ACL SETUSER default -@dangerous

# use a firewall / security group to restrict access
# Redis should NEVER be directly internet-accessible
# (the 2018+ ransom attacks scanned for open Redis instances)

# log slow commands (slowlog)
CONFIG SET slowlog-log-slower-than 10000   # 10ms in microseconds
CONFIG SET slowlog-max-len 128             # keep 128 entries
SLOWLOG GET 10                             # last 10 slow commands
SLOWLOG RESET

# log to a file (redis.conf)
logfile /var/log/redis/redis.log
loglevel notice    # debug|verbose|notice|warning

# monitor all commands (debugging only — huge perf impact)
MONITOR

# latency monitoring
CONFIG SET latency-monitor-threshold 100   # 100ms
LATENCY HISTORY event-name
LATENCY DOCTOR                              # analysis report

# client list (who's connected)
CLIENT LIST
CLIENT GETNAME
CLIENT SETNAME "my-app"

# 1. NEVER expose Redis to the public internet
bind 127.0.0.1
protected-mode yes

# 2. Use strong passwords or ACLs
requirepass "<64-char-random-string>"
# or ACLs with least privilege

# 3. Enable TLS for transit encryption
tls-port 6379

# 4. Restrict dangerous commands
ACL SETUSER default -@dangerous -flushall -keys

# 5. Use a firewall / security group
# allow only app servers on the Redis port

# 6. Rotate credentials regularly
ACL SETUSER appuser on >newpassword ~app:* +@read +@write

# 7. Audit access
LOGICALDB MONITOR  # or use an audit module

# 8. Encrypt sensitive data at rest (disk encryption)
# RDB/AOF files can contain sensitive data

# 9. Keep Redis updated (security patches)
# 10. Isolate environments (separate dev/staging/prod)

# Redis is single-threaded (mostly) — avoid blocking commands
# DANGEROUS on large datasets (block the server):
KEYS *              # use SCAN instead
SMEMBERS bigset     # use SSCAN
HGETALL bighash     # use HSCAN
DEL hugekey         # use UNLINK (async delete)
FLUSHDB             # use FLUSHDB ASYNC

# use UNLINK instead of DEL for large keys (non-blocking)
UNLINK bigkey       # deletes in a background thread

# favor pipelining and batch commands
MGET k1 k2 k3       # one round-trip, not three GETs
LPUSH list a b c    # one command, not three

# prefer structurally efficient commands
# O(1): GET, SET, HGET, SISMEMBER, ZSCORE
# O(log N): ZADD, ZRANGE
# O(N): LRANGE, SMEMBERS, KEYS, SORT

# pipelining: send many commands without waiting for replies
# (reduces round-trip time — huge speedup over network)

# example (Node.js with ioredis):
const pipeline = redis.pipeline()
for (let i = 0; i < 1000; i++) {
  pipeline.set(`key:${i}`, "value")
}
await pipeline.exec()        # one round-trip for 1000 commands

# transactions (MULTI/EXEC) are automatically pipelined
redis.multi().set("k", "v").incr("c").exec()

# MSET/MGET batch in one command
MSET k1 v1 k2 v2 k3 v3
MGET k1 k2 k3

# benchmark pipelining
redis-benchmark -t set -n 100000 -P 16   # -P = pipeline size

# don't pipeline too many at once (memory for queued replies)
# batch in groups of 100-1000

# use the right data structure (memory vs speed trade-offs)
# store small objects as hashes (listpack encoding is compact)
HSET user:1 name "Alice" age 30   # better than 3 string keys

# tune encoding thresholds (redis.conf)
hash-max-listpack-entries 128     # hash -> hashtable above this
hash-max-listpack-value 64        # field value size limit
list-max-listpack-size -2         # listpack max size
set-max-intset-entries 512        # intset -> hashtable above this
zset-max-listpack-entries 128

# use hashes for many small keys (key bucketing)
# instead of SET user:1:clicks 5; SET user:2:clicks 3
# use HSET clicks user:1 5 user:2 3

# monitor memory
INFO memory | grep used_memory_human
MEMORY STATS

# set maxmemory and an eviction policy
maxmemory 4gb
maxmemory-policy allkeys-lru

# when maxmemory is reached, Redis evicts keys per the policy
# maxmemory-policy options:

# no eviction (writes fail with OOM error)
maxmemory-policy noeviction

# LRU (Least Recently Used) — approximate LRU
allkeys-lru         # evict any key (cache use case)
volatile-lru        # only evict keys with TTL set

# LFU (Least Frequently Used) — better for skewed access
allkeys-lfu
volatile-lfu

# random eviction
allkeys-random
volatile-random

# TTL-based (evict soonest-expiring)
volatile-ttl

# for a cache: allkeys-lru or allkeys-lfu (best hit rate)
# for a database: noeviction (never lose data silently)
# mixed (some persistent, some cache): volatile-lru

# redis-benchmark: built-in performance testing
redis-benchmark -t set,get -n 100000 -c 50
# -t: test commands, -n: number of requests, -c: concurrent clients

# test with pipelining
redis-benchmark -t set -n 100000 -c 50 -P 16

# test a specific command with a specific key
redis-benchmark -t get -n 100000 -r 100000 -q
# -r: random keys (INSERT random keys for testing)

# latency check (100 requests, shows distribution)
redis-cli --latency
redis-cli --latency-history    # rolling update
redis-cli --latency-dist       # histogram

# in-memory test (no network overhead)
redis-benchmark -t set -n 100000 -q --threads 4

# monitor while benchmarking
redis-cli INFO stats | grep instantaneous_ops_per_sec

# comprehensive server information
INFO                  # all sections
INFO server           # version, uptime, config
INFO clients          # connected clients
INFO memory           # memory usage, fragmentation
INFO persistence      # RDB/AOF status
INFO stats            # ops/sec, hit rate, keyspace
INFO replication      # primary/replica state
INFO keyspace         # keys per database

# real-time stats
LATENCY DOCTOR        # latency analysis
MEMORY DOCTOR         # memory analysis

# live monitoring (subscribe to a special channel)
SUBSCRIBE __redis__:invalidate

# slow query log
SLOWLOG GET 10
SLOWLOG LEN

# client connections
CLIENT LIST
CLIENT KILL ID <id>   # disconnect a client

# list all connected clients
CLIENT LIST
# id, addr, name, age, idle, flags, db, cmd

# name your connections (for identification)
CLIENT SETNAME "my-app-worker-1"

# get the current client's info
CLIENT INFO

# get the current client's ID
CLIENT ID

# kill a client
CLIENT KILL ADDR 1.2.3.4:12345
CLIENT KILL ID 42
CLIENT KILL TYPE normal    # kill all normal clients

# pause clients (for maintenance, 10 seconds)
CLIENT PAUSE 10000

# set a timeout for idle clients
CONFIG SET timeout 300    # 5 minutes

# max clients
CONFIG SET maxclients 10000

# view config (runtime changes with CONFIG SET)
CONFIG GET maxmemory
CONFIG GET maxmemory-policy
CONFIG GET save
CONFIG GET appendonly

# change config at runtime
CONFIG SET maxmemory 4gb
CONFIG SET maxmemory-policy allkeys-lru
CONFIG SET slowlog-log-slower-than 10000

# persist config changes to redis.conf
CONFIG REWRITE

# view all config (large output)
CONFIG GET *

# typical redis.conf (production):
# bind 127.0.0.1
# protected-mode yes
# port 6379
# maxmemory 4gb
# maxmemory-policy allkeys-lru
# save 3600 1
# appendonly yes
# appendfsync everysec
# tcp-keepalive 300
# timeout 0

# check if a key exists and its type
EXISTS mykey
TYPE mykey
OBJECT ENCODING mykey
OBJECT IDLETIME mykey     # seconds since last access

# memory usage per key
MEMORY USAGE mykey

# find keys blocking the server (DEBUG SLEEP for testing)
DEBUG SLEEP 2             # blocks for 2 seconds (testing only)

# check the latency
LATENCY HISTORY event
LATENCY GRAPH event

# object frequency (LFU mode)
OBJECT FREQ mykey

# inspect a key's internal structure (debugging)
DEBUG OBJECT mykey        # raw internal info

# last saved RDB timestamp
LASTSAVE

# active defragmentation (4.0+)
CONFIG SET activedefrag yes
MEMORY MALLOC-STATS

# upgrade a standalone Redis:
# 1. save the dataset
SAVE
# 2. stop Redis
redis-cli SHUTDOWN
# 3. replace the binary
# 4. start the new version
redis-server /etc/redis/redis.conf

# upgrade a primary/replica pair (zero downtime):
# 1. upgrade the REPLICA first
SHUTDOWN
# 2. start new version on replica
# 3. promote replica to primary (SENTINEL FAILOVER)
# 4. upgrade the old primary (now a replica)
# 5. failback if desired

# upgrade a cluster:
# 1. upgrade one node at a time (rolling)
# 2. for each node: CLUSTER FAILOVER (if primary), upgrade, restart

# always back up before upgrading
# check the release notes for breaking changes
# test the upgrade in staging first

# rolling restart of a cluster (script)
for node in $nodes; do
  redis-cli -h $node CLUSTER FAILOVER
  redis-cli -h $node SHUTDOWN
  ssh $node "systemctl start redis"
done

# cache-aside: app checks cache, falls back to DB, fills cache
# pseudo-code:
function get_user(id):
  # 1. check cache
  data = GET("user:" + id)
  if data:
    return JSON.parse(data)

  # 2. cache miss -> query database
  data = db.query("SELECT * FROM users WHERE id = ?", id)

  # 3. fill cache (with TTL to avoid stale data)
  SET("user:" + id, JSON.stringify(data), "EX", 3600)

  return data

# write-through: update cache when DB changes
function update_user(id, data):
  db.update(id, data)
  SET("user:" + id, JSON.stringify(data), "EX", 3600)

# cache invalidation:
DEL("user:" + id)   # when the DB record is deleted

# SET NX with expiry = simple distributed lock
SET lock:resource "token-abc" NX EX 10
# returns OK if acquired, nil if already held

# release: only delete if the value matches (avoid releasing others' locks)
# (requires a Lua script for atomicity)
EVAL "
  if redis.call('GET', KEYS[1]) == ARGV[1] then
    return redis.call('DEL', KEYS[1])
  end
  return 0
" 1 lock:resource "token-abc"

# Redlock algorithm (multi-instance, higher safety):
# 1. get the time
# 2. SET NX on N (usually 5) independent Redis instances
# 3. if acquired on majority (3/5) within validity window -> locked
# 4. otherwise, release on all instances and retry

# lock options:
#   EX 10        # expire after 10 seconds (safety)
#   NX           # only if not exists
#   "token"      # unique value per lock holder (for safe release)

# fixed window rate limiter (simple)
# INCR + EXPIRE on first request
count = INCR("rate:user:1")
if count == 1:
  EXPIRE("rate:user:1", 60)   # 60-second window
if count > 100:
  return "Rate limited"

# sliding window rate limiter (sorted set)
now = current_time_millis()
ZADD("rate:user:1", now, now)         # add current request
ZREMRANGEBYSCORE("rate:user:1", 0, now - 60000)  # remove old
EXPIRE("rate:user:1", 60)
count = ZCARD("rate:user:1")          # current count
if count > 100:
  return "Rate limited"

# token bucket (Lua script for atomicity)
# maintains a bucket of tokens refilled over time
# see: https://redis.io/docs/patterns/distributed-locks/

# GCRA (Generic Cell Rate Algorithm) — elegant single-key version

# sorted set = perfect for leaderboards
ZADD leaderboard 100 "alice" 250 "bob" 150 "carol"

# top 10 players (highest scores)
ZREVRANGE leaderboard 0 9 WITHSCORES

# a player's rank
ZREVRANK leaderboard "alice"     # 0-indexed from top

# update a score (atomic)
ZINCRBY leaderboard 50 "alice"

# players within a score range (e.g., silver tier 100-199)
ZRANGEBYSCORE leaderboard 100 199

# percentile rank
total = ZCARD(leaderboard)
rank = ZREVRANK(leaderboard "alice"
percentile = (total - rank) / total * 100

# pagination (page 3, 10 per page)
ZREVRANGE leaderboard 20 29 WITHSCORES

# time-bucketed leaderboards (weekly)
ZADD leaderboard:2025-W01 100 "alice"
EXPIRE leaderboard:2025-W01 604800   # 1 week TTL

# Redis is ideal for web sessions (fast, expiring, shared)
# store session as a hash with TTL
HSET session:abc123 user_id 1 last_active 1700000000 data "..."
EXPIRE session:abc123 1800     # 30-minute timeout

# update session (and refresh TTL)
HSET session:abc123 last_active 1700001000
EXPIRE session:abc123 1800     # sliding expiration

# destroy a session
DEL session:abc123

# list active sessions for a user (index by user)
SADD user:1:sessions "session:abc123"
SMEMBERS user:1:sessions       # all session IDs

# cluster-friendly: use a hash tag for the user's sessions
SADD {user:1}:sessions "session:abc123"
# enables multi-key operations on the user's session data

# rotate session ID on privilege change (security)
RENAME session:old session:new
EXPIRE session:new 1800

# run a Lua script inline
EVAL "return redis.call('SET', KEYS[1], ARGV[1])" 1 mykey "hello"
# 1 = number of keys
# mykey = KEYS[1]
# "hello" = ARGV[1]

# return types: string/integer/boolean/table (nil/true/false)
EVAL "return 42" 0
EVAL "return {1, 2, 3}" 0
EVAL "return redis.call('GET', KEYS[1])" 1 mykey

# redis.call() raises error if command fails (stops script)
# redis.pcall() returns error as a table (can handle)
EVAL "local ok, err = pcall(redis.call, 'GET', 'missing'); return err" 0

# load script once, get its SHA1
SCRIPT LOAD "return redis.call('GET', KEYS[1])"
# returns: "fa00..." (the SHA1)

# execute by hash (saves bandwidth on repeated calls)
EVALSHA "fa00..." 1 mykey

# check if a script is cached
SCRIPT EXISTS "fa00..." "abcd..."
# returns: 1) 1  2) 0

# flush the script cache
SCRIPT FLUSH

# list all running scripts (NOSCRIPT errors = cache miss -> reload)
SCRIPT DEBUG SYNC   # blocking debug mode
SCRIPT DEBUG NO     # disable

# atomic compare-and-set (only update if value matches)
local cur = redis.call('GET', KEYS[1])
if cur == ARGV[1] then
  return redis.call('SET', KEYS[1], ARGV[2])
end
return 0

# atomic 'decrement if positive' (semaphore / rate limit)
local val = tonumber(redis.call('GET', KEYS[1]) or 0)
if val > 0 then
  redis.call('DECR', KEYS[1])
  return 1
end
return 0

# atomic list pop + push (move between queues)
local item = redis.call('RPOP', KEYS[1])
if item then
  redis.call('LPUSH', KEYS[2], item)
end
return item

# register a library of functions
FUNCTION LOAD "#!lua name=mylib
redis.register_function('myset', function(keys, args)
  return redis.call('SET', keys[1], args[1])
end)

redis.register_function('double', function(keys, args)
  local v = tonumber(redis.call('GET', keys[1]))
  return redis.call('SET', keys[1], v * 2)
end)
"

# call a function
FCALL myset 1 mykey "hello"
FCALL double 1 counter

# list libraries
FUNCTION LIST
FUNCTION LIST WITHCODE

# delete a library
FUNCTION DELETE mylib

# dump and restore (for backup / migration)
FUNCTION DUMP
FUNCTION RESTORE <payload>

# LOG writes to the Redis log (visible in redis-cli)
EVAL "redis.log(redis.LOG_NOTICE, 'hello from lua'); return 1" 0
# levels: LOG_DEBUG, LOG_VERBOSE, LOG_NOTICE, LOG_WARNING

# return the error object on pcall failure
EVAL "local ok, err = pcall(function()
  error('something broke')
end)
return {ok=tostring(ok), err=tostring(err)}" 0

# debugger (Redis 5.0+)
SCRIPT DEBUG SYNC   # synchronous, blocking
TDEBUG              # step / continue / print
# use redis-cli --ldb to launch debugger:
#   redis-cli --ldb --eval myscript.lua key1 key2 , arg1 arg2

# print variables in the debugger
print myvar
step
continue

# INFO: server, clients, memory, stats, replication, cpu
INFO                # everything
INFO memory         # memory section only
INFO replication    # replica lag, master link
INFO stats          # ops/sec, keyspace hits/misses

# key metrics to watch:
#   used_memory_rss        # RSS from OS (real footprint)
#   mem_fragmentation_ratio # >1.5 = fragmentation, <1 = swapping (bad!)
#   connected_clients       # client connections
#   instantaneous_ops_per_sec
#   keyspace_misses         # cache hit ratio = hits / (hits + misses)

# LATENCY: samples slow events
LATENCY HISTORY event-name
LATENCY DOCTOR      # human-readable diagnosis
LATENCY RESET

# LATENCY GRAPH event-name (ASCII art)

# SLOWLOG: record slow commands (over slowlog-log-slower-than, default 10ms)
CONFIG SET slowlog-log-slower-than 10000   # 10ms in microseconds
CONFIG SET slowlog-max-len 128             # keep last 128 entries

SLOWLOG GET          # last 10 slow commands
SLOWLOG GET 5        # last 5
SLOWLOG RESET        # clear

# each entry: id, timestamp, duration (us), command args, client, addr

# CLIENT LIST: see all connected clients
CLIENT LIST
# fields: id, addr, fd, name, age, idle, flags, db, cmd, ...

# kill a stuck client
CLIENT KILL ADDR 10.0.0.1:12345
CLIENT KILL ID 42

# name your connections for easier debugging
CLIENT SETNAME my-app-worker

# MEMORY USAGE: bytes used by a single key
MEMORY USAGE key
MEMORY USAGE bigset SAMPLES 0   # scan all elements (accurate)

# MEMORY STATS: detailed allocator stats
MEMORY STATS

# MEMORY DOCTOR: diagnosis
MEMORY DOCTOR

# OBJECT ENCODING: see internal representation
OBJECT ENCODING mylist      # quicklist / ziplist / listpack / hashtable
OBJECT REFCOUNT mykey
OBJECT IDLETIME mykey       # seconds since last access
OBJECT FREQ mykey           # LFU access frequency (only with maxmemory-policy LFU)

# why is my key so big? use DEBUG OBJECT (careful, can be slow)
DEBUG OBJECT mykey

# find big keys safely (production-safe, samples)
redis-cli --bigkeys
# scans keyspace, reports largest of each type

# more accurate: --memkeys (memory usage per key)
redis-cli --memkeys
redis-cli --hotkeys      # with LFU policy, finds frequently-accessed keys

# SCAN: iterate keys without blocking (never use KEYS *)
SCAN 0 MATCH user:* COUNT 100
# returns: next-cursor, [key1, key2, ...]
SCAN <next-cursor> ...
# cursor 0 = done

# scan a hash field-by-field
HSCAN myhash 0 COUNT 100
SSCAN myset 0 MATCH prefix*
ZSCAN myzset 0

# inspect a key's TTL (left to live)
TTL mykey           # seconds (-1 = no expire, -2 = no key)
PTTL mykey          # milliseconds

# cluster node status
CLUSTER NODES
# fields: id, addr, flags (master/slave/fail?), slots, ping/pong

# cluster-wide health
CLUSTER INFO
# cluster_state:ok / cluster_slots_assigned / cluster_slots_ok
# cluster_known_nodes

# slot ownership (which node owns a key)
CLUSTER KEYSLOT mykey    # hash slot number
CLUSTER COUNTKEYSINSLOT 1234

# reshard slots (use redis-cli --cluster reshard)
redis-cli --cluster reshard 127.0.0.1:7000 \
  --cluster-from <node-id> --cluster-to <node-id> \
  --cluster-slots 1000 --cluster-yes

# manual failover (on a replica)
CLUSTER FAILOVER          # wait for sync
CLUSTER FAILOVER FORCE    # skip sync (faster, slight data loss risk)
CLUSTER FAILOVER TAKEOVER # no consensus (emergency only)

# verify cluster after changes
redis-cli --cluster check 127.0.0.1:7000
redis-cli --cluster fix 127.0.0.1:7000    # auto-repair minor issues