Redis & Advanced Caching
Monitoring and Debugging Redis
Monitoring and Debugging Redis
Effective monitoring and debugging are crucial for maintaining Redis performance, identifying issues before they impact users, and optimizing resource usage in production environments.
INFO Command - Comprehensive Server Statistics
The INFO command provides detailed information about Redis server status, organized into sections:
Basic INFO Usage:
# Get all information
INFO
# Get specific section
INFO server
INFO clients
INFO memory
INFO persistence
INFO stats
INFO replication
INFO cpu
INFO commandstats
INFO cluster
INFO keyspace
# Get all information
INFO
# Get specific section
INFO server
INFO clients
INFO memory
INFO persistence
INFO stats
INFO replication
INFO cpu
INFO commandstats
INFO cluster
INFO keyspace
Key Metrics to Monitor:
# Server section
redis_version:7.0.5
uptime_in_seconds:864000
uptime_in_days:10
# Clients section
connected_clients:42
blocked_clients:0
tracking_clients:0
# Memory section
used_memory_human:2.50M
used_memory_rss_human:8.20M
mem_fragmentation_ratio:1.23
maxmemory_human:4.00G
# Stats section
total_connections_received:10523
total_commands_processed:1045234
instantaneous_ops_per_sec:1250
keyspace_hits:984521
keyspace_misses:60713
evicted_keys:0
expired_keys:12453
# Server section
redis_version:7.0.5
uptime_in_seconds:864000
uptime_in_days:10
# Clients section
connected_clients:42
blocked_clients:0
tracking_clients:0
# Memory section
used_memory_human:2.50M
used_memory_rss_human:8.20M
mem_fragmentation_ratio:1.23
maxmemory_human:4.00G
# Stats section
total_connections_received:10523
total_commands_processed:1045234
instantaneous_ops_per_sec:1250
keyspace_hits:984521
keyspace_misses:60713
evicted_keys:0
expired_keys:12453
Critical Metrics:
- mem_fragmentation_ratio: Should be 1.0-1.5 (higher = wasted memory)
- keyspace_hits/misses: Hit ratio = hits/(hits+misses), aim for >80%
- evicted_keys: Non-zero = memory pressure, consider increasing maxmemory
- connected_clients: Monitor for connection leaks
- instantaneous_ops_per_sec: Current throughput
MONITOR Command - Real-time Command Logging
MONITOR shows every command processed by Redis in real-time, useful for debugging but impacts performance:
Start Monitoring:
redis-cli MONITOR
Example Output:
1634567890.123456 [0 127.0.0.1:54321] "GET" "user:1000:name"
1634567890.234567 [0 127.0.0.1:54322] "SET" "session:abc" "data"
1634567890.345678 [0 192.168.1.50:12345] "HGETALL" "cart:5000"
1634567890.456789 [0 127.0.0.1:54321] "INCR" "counter:pageviews"
redis-cli MONITOR
Example Output:
1634567890.123456 [0 127.0.0.1:54321] "GET" "user:1000:name"
1634567890.234567 [0 127.0.0.1:54322] "SET" "session:abc" "data"
1634567890.345678 [0 192.168.1.50:12345] "HGETALL" "cart:5000"
1634567890.456789 [0 127.0.0.1:54321] "INCR" "counter:pageviews"
Performance Warning: MONITOR significantly reduces Redis throughput (up to 50%). Only use for short debugging sessions, never in production under heavy load.
Filter MONITOR Output:
# Monitor only SET commands
redis-cli MONITOR | grep SET
# Monitor commands from specific IP
redis-cli MONITOR | grep "192.168.1.50"
# Monitor commands on specific keys
redis-cli MONITOR | grep "user:*"
# Save to file for analysis
redis-cli MONITOR > redis-monitor.log
# Monitor only SET commands
redis-cli MONITOR | grep SET
# Monitor commands from specific IP
redis-cli MONITOR | grep "192.168.1.50"
# Monitor commands on specific keys
redis-cli MONITOR | grep "user:*"
# Save to file for analysis
redis-cli MONITOR > redis-monitor.log
SLOWLOG - Slow Query Analysis
SLOWLOG records queries that exceed a configured threshold, essential for identifying performance bottlenecks:
Configure Slow Log:
# Set threshold to 10ms (10000 microseconds)
CONFIG SET slowlog-log-slower-than 10000
# Keep last 128 slow queries
CONFIG SET slowlog-max-len 128
# Make permanent in redis.conf
slowlog-log-slower-than 10000
slowlog-max-len 128
# Set threshold to 10ms (10000 microseconds)
CONFIG SET slowlog-log-slower-than 10000
# Keep last 128 slow queries
CONFIG SET slowlog-max-len 128
# Make permanent in redis.conf
slowlog-log-slower-than 10000
slowlog-max-len 128
View Slow Log:
# Get last 10 slow queries
SLOWLOG GET 10
Output Format:
1) 1) (integer) 14 # Unique ID
2) (integer) 1634567890 # Unix timestamp
3) (integer) 23451 # Execution time (microseconds)
4) 1) "KEYS" # Command
2) "user:*" # Arguments
5) "127.0.0.1:54321" # Client IP:port
6) "" # Client name
# Get number of entries in log
SLOWLOG LEN
# Clear slow log
SLOWLOG RESET
# Get last 10 slow queries
SLOWLOG GET 10
Output Format:
1) 1) (integer) 14 # Unique ID
2) (integer) 1634567890 # Unix timestamp
3) (integer) 23451 # Execution time (microseconds)
4) 1) "KEYS" # Command
2) "user:*" # Arguments
5) "127.0.0.1:54321" # Client IP:port
6) "" # Client name
# Get number of entries in log
SLOWLOG LEN
# Clear slow log
SLOWLOG RESET
Analysis Tip: Common slow commands include KEYS (use SCAN instead), SORT without LIMIT, and operations on large collections (lists with millions of items).
redis-cli Diagnostic Tools
redis-cli includes built-in tools for performance analysis and diagnostics:
--bigkeys - Find Largest Keys:
redis-cli --bigkeys
Sample Output:
Scanning the entire keyspace...
Biggest string: "cache:homepage" (2048576 bytes)
Biggest list: "queue:emails" (50000 items)
Biggest set: "tags:all" (8934 members)
Biggest hash: "user:1000:profile" (1000 fields)
Biggest zset: "leaderboard:global" (100000 members)
# Options
--bigkeys --i 0.1 # Sample every 0.1 seconds
--bigkeys -a mypass # With authentication
redis-cli --bigkeys
Sample Output:
Scanning the entire keyspace...
Biggest string: "cache:homepage" (2048576 bytes)
Biggest list: "queue:emails" (50000 items)
Biggest set: "tags:all" (8934 members)
Biggest hash: "user:1000:profile" (1000 fields)
Biggest zset: "leaderboard:global" (100000 members)
# Options
--bigkeys --i 0.1 # Sample every 0.1 seconds
--bigkeys -a mypass # With authentication
--latency - Measure Latency:
# Basic latency test
redis-cli --latency
# Output: min: 0, max: 2, avg: 0.15 (1234 samples)
# Latency history (tracks over time)
redis-cli --latency-history
# Latency distribution
redis-cli --latency-dist
# Intrinsic latency (system delays)
redis-cli --intrinsic-latency 100
# Basic latency test
redis-cli --latency
# Output: min: 0, max: 2, avg: 0.15 (1234 samples)
# Latency history (tracks over time)
redis-cli --latency-history
# Latency distribution
redis-cli --latency-dist
# Intrinsic latency (system delays)
redis-cli --intrinsic-latency 100
--stat - Live Statistics:
redis-cli --stat
Output (updates every second):
------- data ------ --------------------- load -------------------- - child -
keys mem clients blocked requests connections
506 1.45M 10 0 10000 (+0) 50
506 1.45M 10 0 10234 (+234) 50
506 1.47M 11 0 10567 (+333) 51
redis-cli --stat
Output (updates every second):
------- data ------ --------------------- load -------------------- - child -
keys mem clients blocked requests connections
506 1.45M 10 0 10000 (+0) 50
506 1.45M 10 0 10234 (+234) 50
506 1.47M 11 0 10567 (+333) 51
--scan - Safe Key Scanning:
# Scan all keys
redis-cli --scan
# Scan with pattern
redis-cli --scan --pattern "user:*"
# Count keys matching pattern
redis-cli --scan --pattern "session:*" | wc -l
# Delete keys matching pattern safely
redis-cli --scan --pattern "temp:*" | xargs redis-cli DEL
# Scan all keys
redis-cli --scan
# Scan with pattern
redis-cli --scan --pattern "user:*"
# Count keys matching pattern
redis-cli --scan --pattern "session:*" | wc -l
# Delete keys matching pattern safely
redis-cli --scan --pattern "temp:*" | xargs redis-cli DEL
Memory Analysis Commands
Detailed memory usage analysis for optimization:
MEMORY Commands:
# Memory usage of specific key
MEMORY USAGE user:1000
# Output: (integer) 1024
# Memory statistics
MEMORY STATS
# Output: Detailed breakdown of memory allocation
# Memory doctor (optimization suggestions)
MEMORY DOCTOR
# Output: Suggestions based on current usage patterns
# Memory usage by sample
MEMORY MALLOC-STATS
# Purge memory (run defragmentation)
MEMORY PURGE
# Memory usage of specific key
MEMORY USAGE user:1000
# Output: (integer) 1024
# Memory statistics
MEMORY STATS
# Output: Detailed breakdown of memory allocation
# Memory doctor (optimization suggestions)
MEMORY DOCTOR
# Output: Suggestions based on current usage patterns
# Memory usage by sample
MEMORY MALLOC-STATS
# Purge memory (run defragmentation)
MEMORY PURGE
Analyze Memory Fragmentation:
INFO memory | grep fragmentation
# mem_fragmentation_ratio:1.23
# If ratio > 1.5, consider:
# 1. Restart Redis to defragment
# 2. Enable active defragmentation
CONFIG SET activedefrag yes
CONFIG SET active-defrag-threshold-lower 10
CONFIG SET active-defrag-threshold-upper 100
INFO memory | grep fragmentation
# mem_fragmentation_ratio:1.23
# If ratio > 1.5, consider:
# 1. Restart Redis to defragment
# 2. Enable active defragmentation
CONFIG SET activedefrag yes
CONFIG SET active-defrag-threshold-lower 10
CONFIG SET active-defrag-threshold-upper 100
CLIENT Commands - Connection Debugging
Monitor and manage client connections:
Client Management:
# List all connected clients
CLIENT LIST
Output Format:
id=123 addr=127.0.0.1:54321 fd=8 name= age=300 idle=5 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=0 obl=0 oll=0 omem=0 events=r cmd=get
# Fields explanation:
# id: Unique client ID
# addr: Client IP and port
# age: Connection age (seconds)
# idle: Idle time (seconds)
# flags: N=normal, M=master, S=slave, P=pubsub
# db: Current database number
# cmd: Last command executed
# List all connected clients
CLIENT LIST
Output Format:
id=123 addr=127.0.0.1:54321 fd=8 name= age=300 idle=5 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=0 obl=0 oll=0 omem=0 events=r cmd=get
# Fields explanation:
# id: Unique client ID
# addr: Client IP and port
# age: Connection age (seconds)
# idle: Idle time (seconds)
# flags: N=normal, M=master, S=slave, P=pubsub
# db: Current database number
# cmd: Last command executed
Client Operations:
# Get current client ID
CLIENT ID
# Set client name
CLIENT SETNAME myapp-web-1
# Get client name
CLIENT GETNAME
# Kill specific client
CLIENT KILL 127.0.0.1:54321
# Kill all clients matching filter
CLIENT KILL TYPE normal SKIPME yes
# Pause all clients for 1000ms
CLIENT PAUSE 1000
# Track client memory
CLIENT TRACKING ON
# Get current client ID
CLIENT ID
# Set client name
CLIENT SETNAME myapp-web-1
# Get client name
CLIENT GETNAME
# Kill specific client
CLIENT KILL 127.0.0.1:54321
# Kill all clients matching filter
CLIENT KILL TYPE normal SKIPME yes
# Pause all clients for 1000ms
CLIENT PAUSE 1000
# Track client memory
CLIENT TRACKING ON
Latency Monitoring
Detect and diagnose latency spikes:
Enable Latency Monitoring:
# Monitor events taking > 100ms
CONFIG SET latency-monitor-threshold 100
# View recent latency events
LATENCY LATEST
Output:
1) 1) "command"
2) (integer) 1634567890
3) (integer) 250
4) (integer) 350
# View latency history for specific event
LATENCY HISTORY command
# Latency doctor (diagnosis)
LATENCY DOCTOR
# Reset latency data
LATENCY RESET
# Monitor events taking > 100ms
CONFIG SET latency-monitor-threshold 100
# View recent latency events
LATENCY LATEST
Output:
1) 1) "command"
2) (integer) 1634567890
3) (integer) 250
4) (integer) 350
# View latency history for specific event
LATENCY HISTORY command
# Latency doctor (diagnosis)
LATENCY DOCTOR
# Reset latency data
LATENCY RESET
Latency Events:
command # Slow commands
fast-command # Fast command taking too long
fork # Fork operation for persistence
aof-write # AOF write operation
aof-fsync-always # AOF sync to disk
rdb-unlink-temp-file # RDB file cleanup
command # Slow commands
fast-command # Fast command taking too long
fork # Fork operation for persistence
aof-write # AOF write operation
aof-fsync-always # AOF sync to disk
rdb-unlink-temp-file # RDB file cleanup
Redis Insight - GUI Monitoring Tool
Redis Insight provides a graphical interface for monitoring, profiling, and debugging:
Redis Insight Features:
- Browser: Visual key browser with search and filters
- Profiler: Real-time command profiling (like MONITOR)
- Slowlog: Visual slow query analysis
- Memory Analysis: Memory usage charts and recommendations
- CLI: Built-in command-line interface
- Cluster Management: Multi-node cluster overview
- Workbench: Execute commands and scripts
Install Redis Insight:
# Download from RedisLabs
https://redis.com/redis-enterprise/redis-insight/
# Or via Docker
docker run -d -p 8001:8001 redislabs/redisinsight:latest
# Access at http://localhost:8001
# Download from RedisLabs
https://redis.com/redis-enterprise/redis-insight/
# Or via Docker
docker run -d -p 8001:8001 redislabs/redisinsight:latest
# Access at http://localhost:8001
Command Statistics
Track which commands are used most frequently:
View Command Stats:
INFO commandstats
Output:
cmdstat_get:calls=458234,usec=1234567,usec_per_call=2.69
cmdstat_set:calls=312456,usec=987654,usec_per_call=3.16
cmdstat_hgetall:calls=12345,usec=456789,usec_per_call=37.01
# Reset command stats
CONFIG RESETSTAT
INFO commandstats
Output:
cmdstat_get:calls=458234,usec=1234567,usec_per_call=2.69
cmdstat_set:calls=312456,usec=987654,usec_per_call=3.16
cmdstat_hgetall:calls=12345,usec=456789,usec_per_call=37.01
# Reset command stats
CONFIG RESETSTAT
Optimization Tip: Commands with high usec_per_call are candidates for optimization. Look for HGETALL on large hashes, SMEMBERS on large sets, or KEYS usage.
Persistence Monitoring
Monitor RDB and AOF operations:
RDB Status:
INFO persistence
rdb_last_save_time:1634567890
rdb_changes_since_last_save:12345
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:5
rdb_current_bgsave_time_sec:-1
AOF Status:
aof_enabled:1
aof_rewrite_in_progress:0
aof_last_rewrite_time_sec:12
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
INFO persistence
rdb_last_save_time:1634567890
rdb_changes_since_last_save:12345
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:5
rdb_current_bgsave_time_sec:-1
AOF Status:
aof_enabled:1
aof_rewrite_in_progress:0
aof_last_rewrite_time_sec:12
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
Debugging Common Issues
Issue: High Memory Usage
1. Run
2. Check
3. Review evicted_keys - if >0, increase maxmemory
4. Check for keys without TTL:
Issue: Slow Performance
1. Check
2. Run
3. Verify
4. Check
Issue: Connection Problems
1. Check
2. Look for rejected_connections in
3. Review timeout setting in redis.conf
4. Check network latency with
1. Run
redis-cli --bigkeys to find large keys2. Check
INFO memory for mem_fragmentation_ratio3. Review evicted_keys - if >0, increase maxmemory
4. Check for keys without TTL:
redis-cli --scan | xargs redis-cli TTL | grep -c "\-1"Issue: Slow Performance
1. Check
SLOWLOG GET 100 for slow queries2. Run
MONITOR briefly to see live commands3. Verify
INFO stats for keyspace_misses ratio4. Check
CLIENT LIST for blocked clientsIssue: Connection Problems
1. Check
INFO clients for connected_clients vs maxclients2. Look for rejected_connections in
INFO stats3. Review timeout setting in redis.conf
4. Check network latency with
redis-cli --latencyMonitoring Best Practices
Production Monitoring Checklist:
- ✅ Set up automated INFO collection every 60 seconds
- ✅ Alert on mem_fragmentation_ratio > 1.5
- ✅ Alert on keyspace hit ratio < 80%
- ✅ Alert on evicted_keys > 0
- ✅ Monitor connected_clients for leaks
- ✅ Review SLOWLOG daily for optimization opportunities
- ✅ Track instantaneous_ops_per_sec for capacity planning
- ✅ Monitor used_memory vs maxmemory
- ✅ Set latency-monitor-threshold and review LATENCY DOCTOR
- ✅ Use Redis Insight for visual troubleshooting
Exercise: Set up comprehensive Redis monitoring: 1) Create a script that runs INFO every 60 seconds and logs key metrics, 2) Enable SLOWLOG with 10ms threshold and analyze slow queries, 3) Use CLIENT LIST to identify idle connections, 4) Run redis-cli --bigkeys to find memory-intensive keys, 5) Check memory fragmentation and enable active defragmentation if needed.