Mastering Redis Monitoring: Key Metrics, Commands, and Performance Testing

Monitoring Metrics Overview

Redis provides a set of built‑in metrics that can be grouped into five categories: Performance, Memory, Basic Activity, Persistence, and Error. These metrics help operators understand the health, capacity, and reliability of a Redis instance.

Performance Metrics

Name: latency – Description: Time taken for Redis to respond to a request.

Name: instantaneous_ops_per_sec – Description: Average number of operations processed per second.

Name: hit_rate (calculated) – Description: Cache hit rate, derived from other counters.

Memory Metrics

Name: used_memory – Description: Total memory allocated by the allocator from the operating system.

Name: mem_fragmentation_ratio – Description: Ratio of memory fragmentation.

Name: evicted_keys – Description: Number of keys evicted because the maxmemory limit was reached.

Name: blocked_clients – Description: Clients blocked by BLPOP, BRPOP, or BRPOPLPUSH.

Basic Activity Metrics

Name: connected_clients – Description: Number of client connections.

Name: connected_slaves – Description: Number of slave connections.

Name: master_last_io_seconds_ago – Description: Seconds since the last master‑slave interaction.

Name: keyspace – Description: Total number of keys stored in the database.

Persistence Metrics

Name: rdb_last_save_time – Description: Unix timestamp of the last RDB snapshot saved to disk.

Name: rdb_changes_since_last_save – Description: Number of changes made to the dataset since the last RDB save.

Error Metrics

Name: rejected_connections – Description: Connections rejected because the maxclients limit was reached.

Name: keyspace_misses – Description: Number of key look‑ups that missed (no hit).

Name: master_link_down_since_seconds – Description: Duration, in seconds, that the master‑slave link has been down.

Monitoring Tools and Commands

redis-benchmark – synthetic benchmark tool.

redis-stat – real‑time statistics collector.

redis-faina – performance analysis utility.

redislive – live monitoring dashboard.

redis-cli – interactive command‑line client.

MONITOR – streams every command processed by the server.

SHOWLOG – displays server logs.

Slow‑query log management:

slowlog-log-slower-than 1000   # Log queries slower than 1000 µs</code>
<code>slowlog-max-len 100          # Keep up to 100 entries in the slowlog

Retrieving information with INFO (full dump or per‑section):

server – server environment parameters.

clients – client‑related information.

memory – memory usage statistics.

persistence – persistence status.

stats – general statistics.

replication – master‑slave replication details.

cpu – CPU usage.

cluster – cluster information.

keyspace – key count per database.

Example commands:

./redis-cli info stats | grep ops   # Show operations per second

# Interactive INFO usage
./redis-cli
> INFO server

Performance Monitoring

To measure raw throughput, use redis-cli info | grep ops or the benchmark tool:

redis-benchmark -c 100 -n 5000   # 100 concurrent connections, 5000 requests

Memory Monitoring

Key memory counters can be extracted with:

# Human‑readable memory usage
./redis-cli info | grep used_memory_human

# Blocked client count
./redis-cli info | grep blocked_clients

# Evicted keys count
./redis-cli info | grep evicted_keys

# Memory fragmentation ratio
./redis-cli info | grep mem_fragmentation_ratio

Basic Activity Monitoring

Client connections and replication status are visible via:

# Connected clients
./redis-cli info | grep connected_clients

# Connected slaves
./redis-cli info | grep connected_slaves

Persistence Monitoring

# Last RDB save timestamp
./redis-cli info | grep rdb_last_save_time

# Changes since last save
./redis-cli info | grep rdb_changes_since_last_save

Error Monitoring

Important error‑related counters:

# Rejected connections due to maxclients
./redis-cli info | grep rejected_connections

# Keyspace misses
./redis-cli info | grep keyspace_misses

# Master link down duration
./redis-cli info | grep master_link_down_since_seconds

# Replication backlog size
./redis-cli info | grep repl_backlog_size

# Partial sync errors
./redis-cli info | grep sync_partial_err

Redis Performance Testing Commands

redis-benchmark -c 100 -n 5000   # 100 connections, 5000 total requests

This command generates a synthetic load to evaluate throughput and latency under controlled conditions.