redis-cli and Debugging

redis-cli                                  # 127.0.0.1:6379
redis-cli -h cache.example.com -p 6380
redis-cli -h host -p 6379 -a "$REDIS_PASSWORD" -n 3      # -n selects the database
redis-cli -u 'redis://user:pass@host:6379/0'             # URI form
redis-cli -u 'rediss://user:pass@host:6380/0' --tls      # TLS
redis-cli --tls --cacert ca.crt --cert client.crt --key client.key -h host -p 6380
redis-cli -c -h host                                     # follow Cluster redirects
redis-cli -3                                             # negotiate RESP3 (HELLO 3)

redis-cli --raw GET hu:k              # no quoting, no type prefixes, UTF-8 through
redis-cli --no-raw LRANGE hu:l 0 -1   # force the pretty numbered form in a pipe
redis-cli --json GET hu:json          # JSON output
redis-cli -x SET hu:blob < photo.bin  # last argument from stdin
redis-cli --quoted-input SET k "a\x00b"    # accept escapes in the input
redis-cli -r 5 -i 1 INFO clients | grep connected_clients
redis-cli --eval script.lua key1 key2 , arg1 arg2   # note the SPACED comma

redis-cli --scan                                  # every key, safely, via SCAN
redis-cli --scan --pattern 'hu:session:*'
redis-cli --scan --pattern 'hu:tmp:*' --count 1000 | xargs -L 100 redis-cli UNLINK

redis-cli --bigkeys        # biggest key per type, by element count
redis-cli --memkeys        # biggest key per type, by MEMORY USAGE bytes
redis-cli --keystats       # both at once, plus a size distribution
redis-cli --hotkeys        # most accessed keys, REQUIRES an LFU maxmemory-policy
redis-cli --bigkeys --pattern 'hu:*'

redis-cli --latency                  # min, max, avg over a rolling window
redis-cli --latency-history -i 5     # a new sample line every 5 seconds
redis-cli --latency-dist             # a spectrum, needs a 256-color terminal
redis-cli --latency-percentiles      # p50, p99, p99.9
redis-cli --intrinsic-latency 10     # measures the KERNEL, not Redis

redis-cli CONFIG SET latency-monitor-threshold 100    # record events over 100ms
redis-cli LATENCY LATEST      # event, last time, last value, max value
redis-cli LATENCY HISTORY command
redis-cli LATENCY RESET
redis-cli LATENCY DOCTOR      # prose analysis of what it recorded
redis-cli MEMORY DOCTOR       # prose analysis of memory

redis-cli CONFIG SET slowlog-log-slower-than 10000   # microseconds, 10000 = 10ms
redis-cli CONFIG SET slowlog-max-len 256
redis-cli SLOWLOG GET 10
redis-cli SLOWLOG LEN
redis-cli SLOWLOG RESET
redis-cli SLOWLOG HELP

1) 1) (integer) 14                     <- entry id
   2) (integer) 1789531888             <- unix timestamp
   3) (integer) 42311                  <- microseconds of EXECUTION time
   4) 1) "HGETALL"                     <- the command and its args, truncated
      2) "hu:user:1042"
   5) "127.0.0.1:61083"                <- client address
   6) ""                               <- client name, from CLIENT SETNAME

redis-cli MONITOR
# 1789531888.747123 [0 127.0.0.1:61083] "HGETALL" "hu:user:1042"
redis-cli MONITOR | grep -i flushall            # what you are usually looking for
redis-cli MONITOR | head -1000 > /tmp/trace.txt

redis-cli INFO                    # every default section
redis-cli INFO memory
redis-cli INFO replication clients        # several sections, 7.0+
redis-cli INFO commandstats
redis-cli INFO latencystats               # 7.0+
redis-cli INFO everything                 # including the expensive sections
redis-cli CONFIG RESETSTAT                # zero the counters

# Cache hit rate
redis-cli INFO stats | grep -E 'keyspace_(hits|misses)'

# The three numbers that explain most incidents
redis-cli INFO memory | grep -E 'used_memory_human|used_memory_peak_human|mem_fragmentation_ratio'
redis-cli INFO stats  | grep -E 'evicted_keys|expired_keys|rejected_connections'
redis-cli INFO clients | grep -E 'connected_clients|blocked_clients'

redis-cli CLIENT LIST                        # every connection, one line each
redis-cli CLIENT LIST TYPE pubsub            # normal | master | replica | pubsub
redis-cli CLIENT INFO                        # just this connection
redis-cli CLIENT ID
redis-cli CLIENT SETNAME importer            # shows up in CLIENT LIST and SLOWLOG
redis-cli CLIENT KILL ID 42
redis-cli CLIENT KILL ADDR 10.0.0.9:51234
redis-cli CLIENT KILL LADDR 10.0.0.1:6379 TYPE normal SKIPME yes
redis-cli CLIENT NO-EVICT on                 # 7.0+, exempt this client from eviction
redis-cli CLIENT NO-TOUCH on                 # 7.2+, do not update LRU/LFU on my reads
redis-cli CLIENT UNPAUSE
redis-cli CLIENT PAUSE 5000 WRITE            # stall writes, for a controlled failover

redis-cli DEBUG OBJECT hu:user:1042
# Value at:0x... refcount:1 encoding:listpack serializedlength:24 lru:... lru_seconds_idle:0
redis-cli DEBUG JMAP                # no-op on most builds
redis-cli DEBUG SLEEP 1             # block the server for 1 second, ON PURPOSE
redis-cli DEBUG SET-ACTIVE-EXPIRE 0 # disable the expiry cycle, for tests
redis-cli DEBUG QUICKLIST-PACKED-THRESHOLD 100

redis-benchmark -h host -p 6379 -q                     # quick summary of every test
redis-benchmark -t get,set -n 100000 -c 50 -P 16 -q    # ops, clients, pipeline depth
redis-benchmark -t set -n 100000 -d 1024 -q            # 1KB values
redis-benchmark -n 10000 -r 100000 -t set -q           # 100k random keyspace
redis-benchmark -n 1000 script load "return 1"
redis-benchmark --dbnum 9 -t set -n 1000               # keep it out of db 0

redis-cli --rdb /tmp/backup.rdb                # ask the server for a fresh RDB, over the wire
redis-cli --functions-rdb /tmp/funcs.rdb       # only the function libraries
redis-cli BGSAVE && redis-cli LASTSAVE         # snapshot in place, confirm it advanced

redis-cli DUMP hu:user:1042 | xxd | head -2    # one key, serialized
redis-cli -h src DUMP hu:k                     # then RESTORE on the target
redis-cli MIGRATE targethost 6379 hu:k 0 5000 COPY REPLACE   # move one key live