Redis Monitoring Metrics and Commands Guide

The article presents a detailed guide on monitoring Redis instances, covering key metric categories, specific metric definitions, recommended monitoring tools, configuration snippets, and practical command-line examples for extracting and analyzing performance data.

Monitoring Metrics

Performance metrics

Memory metrics

Basic activity metrics

Persistence metrics

Error metrics

Performance Metrics

Name

Description

latency

Redis response time for a request

instantaneous_ops_per_sec

Average number of operations processed per second

hi rate(calculated)

Cache hit rate (calculated)

Memory Metrics

Name

Description

used_memory

Memory currently used

mem_fragmentation_ratio

Memory fragmentation ratio

evicted_keys

Number of keys evicted due to max‑memory limit

blocked_clients

Clients blocked by BLPOP/BRPOP/BRPOPLPUSH

Basic Activity Metrics

Name

Description

connected_clients

Number of client connections

conected_laves

Number of slave replicas

master_last_io_seconds_ago

Seconds since last master‑slave interaction

keyspace

Total number of keys in the database

Persistence Metrics

Name

Description

rdb_last_save_time

Timestamp of the last RDB persistence

rdb_changes_sice_last_save

Number of changes since the last persistence

Error Metrics

Name

Description

rejected_connections

Connections rejected due to maxclient limit

keyspace_misses

Number of key look‑ups that missed

master_link_down_since_seconds

Duration (seconds) of master‑slave link down

Monitoring Tools

redis-benchmark

redis-stat

redis-faina

redislive

redis-cli

monitor

showlog

Typical usage examples include retrieving slow‑log entries, checking configuration via info, and filtering specific sections with grep:

slowlog-log-slower-than 1000   # set slow‑query threshold (µs)
slowlog-max-len 100           # set max number of slow‑log entries

Example commands for extracting metrics:

# Get operations per second
redis-cli info | grep ops

# Memory usage details
redis-cli info | grep used | grep human

# Blocked clients
redis-cli info | grep blocked_clients

Performance Testing

./redis-benchmark -c 100 -n 5000

This runs a benchmark with 100 concurrent connections and 5,000 total requests, useful for assessing Redis throughput.

Overall, the guide equips operators with the necessary metrics, commands, and tools to monitor Redis health, diagnose issues, and optimize performance.