Redis monitoring.

Key metrics

  • Connected clients.
  • Ops/sec.
  • Memory used.
  • Eviction count.
  • Hit/miss ratio.
  • Replication lag.
  • Latency (p99).

INFO

redis-cli INFO clients
redis-cli INFO memory
redis-cli INFO stats
redis-cli INFO replication
redis-cli INFO commandstats
redis-cli INFO latencystats

SLOWLOG

CONFIG SET slowlog-log-slower-than 10000   # 10ms
SLOWLOG GET 10
SLOWLOG LEN
SLOWLOG RESET

Latency

redis-cli --latency
redis-cli --latency-history -i 5
redis-cli --latency-dist
redis-cli LATENCY DOCTOR

Commandstats

redis-cli INFO commandstats
# cmdstat_set:calls=1234,usec=5678,usec_per_call=4.6

Hot keys (LFU)

CONFIG SET maxmemory-policy allkeys-lfu
redis-cli --hotkeys

Big keys

redis-cli --bigkeys
redis-cli --memkeys

redis-exporter (Prometheus)

docker run -d --name redis-exporter -p 9121:9121 \
    -e REDIS_ADDR=redis://redis:6379 \
    oliver006/redis_exporter

Exposes metrics at :9121/metrics. Grafana dashboards: 763, 11835.

Useful PromQL

# Hit rate
rate(redis_keyspace_hits_total[5m])
  / (rate(redis_keyspace_hits_total[5m]) + rate(redis_keyspace_misses_total[5m]))

# Evictions
rate(redis_evicted_keys_total[5m])

# Memory
redis_memory_used_bytes / redis_memory_max_bytes

# Replication lag (seconds)
redis_slave_lag

Alerts

  • redis_up == 0 for 1m.
  • evictions/sec > 0 for 5m.
  • memory_used / maxmemory > 0.9.
  • connected_clients > 10000 (or your max).
  • replication_lag > 10.

Healthcheck

redis-cli ping
# PONG

CONFIG

CONFIG GET maxmemory*
CONFIG GET save
CONFIG SET maxmemory 4gb
CONFIG REWRITE        # save to file

Latency events

LATENCY LATEST
LATENCY HISTORY event-name
LATENCY RESET

Events: expire-cycle, eviction-cycle, fork, etc.

Persistence latency

Heavy BGSAVE forks → latency spikes. Watch LATENCY LATEST for fork events.

Active expiration

If TTLs heavy: tune hz (default 10/sec).

CONFIG SET hz 100

Higher hz → more CPU spent expiring, less memory.

Hot patterns

# What commands are slow?
SLOWLOG GET 100 | grep -E 'KEYS|SMEMBERS|HGETALL'

# Big sets/lists?
redis-cli --bigkeys

KEYS, SMEMBERS, HGETALL, ZRANGE 0 -1 on huge keys block server.

Capacity planning

  • Avg key size × keys = memory.
  • Plus 20-30% overhead.
  • Plus headroom for fork (snapshot doubles memory briefly).

Reduce memory footprint

  • Compact encodings (small structures).
  • Shorter keys / values.
  • Compression (zstd) for big values.
  • Use HyperLogLog instead of sets where applicable.

Common mistakes

  • Monitoring CPU but not memory.
  • Slowlog disabled → can’t debug.
  • No alerts on evictions.
  • Forgetting BGSAVE fork costs RAM.
  • Single Redis for cache + queue + sessions → noisy neighbor.

Read this next

If you want my Redis monitoring stack, it’s at rajpoot.dev .

Building something AI-, backend-, or data-heavy and want a second pair of eyes? I do consulting and freelance work — see my projects and ways to reach me at rajpoot.dev .