Redis configuration file best practices

# Binded IP address (default binding 127.0.0.1, only local access)# It is recommended to bind the specific IP or 0.0.0.0 in the production environment (need to be accompanied by a firewall/password)bind 0.0.0.0
# Listen to port (default 6379)port 6379
# Whether to run in daemon mode (default no), it is recommended to use yes in the production environmentdaemonize yes
# PID file pathpidfile /var/run/redis_6379.pid
# Log level (debug|verbose|Notice|warning, default notice)loglevel notice
# Log file path (set to stdout output to console, default is empty)logfile "/var/log/redis/"
# Number of databases (0-15, default 16)databases 16
# Whether to enable protected mode (default yes, only local access or password access is allowed)protected-mode yes
# Maximum number of client connections (default 10000, set to 0 to indicate no limit)maxclients 10000

# Maximum memory limit (such as 2gb, 512mb, default 0 means no limit, must be set in the production environment)maxmemory 0
# Memory elimination strategy (default noeviction)# - noeviction: prohibit writing, return an error# - allkeys-lru: Remove the least recently used key# - allkeys-random: randomly remove keys# - volatile-lru: Remove the LRU key with expiration time set# - volatile-random: Randomly remove key with expiration time set# - volatile-ttl: Remove the key that is about to expire# - volatile-lfu: Remove the least frequently used key with expiration time# - allkeys-lfu: Remove the least frequently used keysmaxmemory-policy noeviction
# Approximate accuracy of LRU/LFU algorithm (the larger the value, the more accurate it is, the default is 50)maxmemory-samples 50
# Whether to enable automatic memory defragmentation (default no)activedefrag no
# Minimum memory fragmentation triggered by defragmentation (default 100mb)active-defrag-ignore-bytes 100mb
# Minimum fragmentation rate triggered by defragmentation (default 10%)active-defrag-threshold-lower 10
# Maximum fragmentation rate triggered by defragmentation (default 100%)active-defrag-threshold-upper 100
# The lower limit of CPU time used for defragmentation (default 1%)active-defrag-cycle-min 1
# The maximum CPU time limit for defragmentation (default 25%, avoid affecting normal services)active-defrag-cycle-max 25

# Conditions for automatically generating RDB snapshots (save <seconds> <changes>)# Multiple rules can be set, and any condition will be triggeredsave 900 1      # At least 1 key is modified within 900 secondssave 300 10     # At least 10 keys are modified within 300 secondssave 60 10000   # At least 10,000 keys are modified within 60 seconds# Disable RDB snapshots (comment all save lines)# save ""
# RDB file name (default)dbfilename 
# RDB file storage path (must be a directory)dir ./
# Whether the RDB file is compressed (default yes, compression may affect performance)rdbcompression yes
# Whether to add CRC64 checksum to RDB file (default yes, sacrificing about 10% performance for data integrity)rdb-checksum yes
# Whether to use incremental fsync when generating RDB in the background (default yes, 4.0+ support)rdb-save-incremental-fsync yes

# Whether to enable AOF persistence (default no)appendonly no
# AOF file name (default)appendfilename ""
# AOF synchronization policy (default everysec)# - always: sync to disk every write operation (slowest but safest)# - everysec: Synchronize once per second (taking into account both performance and security)# - no: It is up to the operating system to decide when to sync (fastest but potentially lose data)appendfsync everysec
# Whether to pause fsync during AOF rewrite (default no, avoid data loss)no-appendfsync-on-rewrite no
# Minimum size for automatic rewrite of AOF files (default 64mb)auto-aof-rewrite-min-size 64mb
# The growth rate of automatic rewriting of AOF files (default 100%, that is, the current file is triggered when 2 times the last rewrite)auto-aof-rewrite-percentage 100
# Whether to ignore the last possible incomplete command when loading the AOF file (default yes)aof-load-truncated yes
# Whether to use RDB format prefix when rewriting AOF (reduce file size, default yes)aof-use-rdb-preamble yes

# Set the current Redis as the slave and specify the master node (master IP and port)replicaof <masterip> <masterport>
# Master Node Password (if any)masterauth <password>
# Is the slave node read-only (default yes)replica-read-only yes
# The network timeout time of master-slave replication (seconds, default 60)repl-timeout 60
# Master-slave replicate heartbeat frequency (seconds, default 10)repl-ping-replica-period 10
# Copy buffer size (used to save write commands that the master node is not synchronized to the slave node, default 1mb)repl-backlog-size 1mb
# The persistence time of the copy buffer (seconds, memory will be released after this time and no node connection is available, default 3600)repl-backlog-ttl 3600
# Slave node priority (the smaller the value, the higher the priority, 0 means not participating in the election, default 100)replica-priority 100
# When the master node is not available, will the slave node continue to serve the read request (default yes)replica-serve-stale-data yes
# Whether to use diskless replication (the master node sends RDB directly through the network to avoid disk I/O, default no)repl-diskless-sync no
#Diskless replication delay time (seconds, wait for more slave nodes to be connected and transmitted together, default 5)repl-diskless-sync-delay 5

# Set the access password (it is recommended to use complex passwords, which must be set in the production environment)requirepass foobared
# Whether to enable ACL (access control list, replace traditional password, default no)aclfile /etc/redis/
# Rename dangerous commands (for example, change FLUSHALL to a safe name, or disable the command)# rename-command FLUSHALL "" # Disable commands# rename-command FLUSHALL "FLUSHDB"  # Rename command（Not recommended）

# Backlog queue length of TCP connection (default 511, affecting burst connection processing capabilities in a short time)tcp-backlog 511
# TCP connection keep alive time (seconds, default 300)tcp-keepalive 300
# Whether to disable TCP_NODELAY (default no, enable Nagle algorithm to reduce the number of network packets, but may increase latency)tcp-nodelay no
# Client idle timeout time (units of seconds, default 0 means no timeout)timeout 0

# Slow query threshold (microseconds, default 10000=10ms)slowlog-log-slower-than 10000
# Maximum length of slow query log (default 128)slowlog-max-len 128

# Whether to enable I/O multi-threading (default no, 4.0+ support, only accelerate network I/O, command execution is still single-threaded)io-threads-do-reads no
# Number of I/O threads (it is recommended to set to half of the number of CPU cores, no more than 8, default 4)io-threads 4

# Whether to enable cluster mode (default no), when set to yes, the Redis instance will run as the cluster node, participating in cluster management and data distribution.cluster-enabled no
# Cluster configuration file (automatically generated, no manual editing required)cluster-config-file 
# Node timeout time (milliseconds, default 15000)cluster-node-timeout 15000
# The minimum length of replication from a node during failover (seconds, default 10)cluster-replica-validity-factor 10
# Does the cluster require all slots to be allocated (default yes to avoid split brain)cluster-require-full-coverage yes
# Can slave nodes be migrated to other master nodes (default yes)cluster-migration-barrier 1

# The upper limit of client connections to publish subscriptions (the default is not limited, setting it to 0 means no limit)pubsub-max-connections 0
# Upper limit of message queue length for publish subscription (default 32, prevent memory overflow)client-output-buffer-limit pubsub 32mb 8mb 60

# Whether to record memory allocation information (default no, recommended to turn off in production environment)memtier yes
# Whether to record memory fragmentation information (default yes)meminfo-command yes
# Whether to record the detailed information of the memory allocator (default no)malloc-stats no

# Whether to check AOF files at startup (default yes)aof-checksum yes
# Whether to enable Lua scripting function (default yes)lua-time-limit 5000  # Maximum execution time of Lua script (milliseconds, default 5000)# Whether to enable progressive rehash of hash tables (default yes, improve performance)activerehashing yes
# Whether to enable large page memory (default no, enabling may cause delay problems)transparent-hugepage yes
# Whether to enable client tracing (default no, used to debug client request sources)client-tracking no
# Redirect mode for client tracking (default OFF)# - ON: Track all clients# - REDIRECT: Redirect trace information to the specified clientclient-tracking-redirection ""
# Whether to enable delay monitoring (default no)latency-monitor-threshold 0  # Set to 0 means disable. Set to a specific value (such as 10) to monitor operations exceeding this value.# The history length of delayed monitoring (default 1024)latency-history-length 1024