Monitoring

Redis MONITOR — Real-Time Command Inspection

MONITOR streams every command that Redis processes in real time.

What it shows:

• Every Redis command executed
• The exact arguments
• The client issuing it
• Millisecond timestamps

When to use:

• Debugging application behavior
• Detecting unexpected commands
• Real-time auditing
• Checking if clients are overloading Redis with unnecessary calls

Caution:

• MONITOR is slow and CPU-heavy because it logs every operation.
• Use on development or temporary debugging, not production long-term.

Output Example:

1617043000.123456 [0 127.0.0.1:51726] "SET" "user:1" "Alice"
1617043000.123789 [0 127.0.0.1:51726] "INCR" "visits"
1617043000.124001 [0 127.0.0.1:51726] "LPUSH" "tasks" "task1"

• Timestamp: 1617043000.123456
• DB index: [0
• Client IP: 127.0.0.1
• Command: "SET", "INCR", "LPUSH"

Redis INFO — Health, Stats, and System Overview

INFO gives a complete snapshot of Redis server status.

It has many sections, including:

Section	Shows
server	Redis version, uptime
clients	Connected clients, blocked clients
memory	RAM usage, fragmentation
persistence	RDB/AOF status
stats	Command statistics
replication	Master/slave info
cpu	CPU usage
cluster	Cluster status
keyspace	Key counts, expirations

Output Example:

# Server
redis_version:7.0.0
uptime_in_seconds:3600

# Clients
connected_clients:42
blocked_clients:1

# Memory
used_memory:25600000
used_memory_peak:30000000
mem_fragmentation_ratio:1.12

# Stats
total_commands_processed:105000
instantaneous_ops_per_sec:1200

# Keyspace
db0:keys=5000,expires=2000,avg_ttl=60000

• Server section
  • Redis has been running for 1 hour
  • Version: 7.0.0
• Clients
  • 42 active connections
  • 1 blocked client (likely due to BRPOP or Lua script)
• Memory
  • used_memory: ~25 MB
  • mem_fragmentation_ratio: 1.12 (healthy)
• Stats
  • 105,000 commands processed overall
  • Throughput: 1200 ops/sec
• Keyspace
  • Database db0 has 5,000 keys
  • 2,000 of them have TTLs

Useful options

• Get only memory info: INFO memory
• Get only clients info: INFO clients

Redis SLOWLOG — Detect Slow Commands

SLOWLOG records Redis commands that exceed a configured execution time threshold.

Redis is fast, so anything > 1 millisecond is often suspicious.

How to configure Slow Log

• Set threshold to 1000 microseconds (1ms): CONFIG SET slowlog-log-slower-than 1000
• Set number of entries stored: CONFIG SET slowlog-max-len 128

Adding a Slow Command

If your application runs:

LRANGE huge_list 0 -1

…on a list with millions of items, it may take >1ms.

Redis will record it in the slow log.

Viewing the Slow Log

SLOWLOG GET 5

Example output:

1) 1) (integer) 15
   2) (integer) 1617043400
   3) (integer) 1200
   4) 1) "LRANGE"
      2) "biglist"
      3) "0"
      4) "-1"
   5) "127.0.0.1:51726"

Field	Meaning
(integer) 15	Slowlog entry ID
1617043400	Unix timestamp
1200	Execution time in microseconds (1.2ms)
"LRANGE" "biglist" "0" "-1"	Command that was slow
"127.0.0.1:51726"	Client that executed it

Redis caught a slow command.

• Get slowlog length: SLOWLOG LEN
• Reset slowlog: SLOWLOG RESET

Memory Usage Inspection

Command	Purpose
INFO memory	High-level memory stats
MEMORY USAGE key	Inspect single key memory
MEMORY STATS	Deep memory metrics
MEMORY DOCTOR	Automated advice
MEMORY MALLOC-STATS	Allocator-level stats
MEMORY PURGE	Defrag attempt