Common Redis Latency Issues and How to Diagnose Them

Redis, as an in‑memory database, can handle up to 100k QPS, but latency spikes often arise from misuse or operational issues.

High‑complexity commands : Commands with O(n) or higher complexity (e.g., SORT, SUNION, ZUNIONSTORE) on large data sets can block the single Redis thread. Enabling the slow‑log (e.g., CONFIG SET slowlog-log-slower-than 5000 and CONFIG SET slowlog-max-len 1000) helps identify such commands.

Large keys : Storing excessively large values increases memory allocation and release time. The redis-cli --bigkeys -i 0.01 command scans key size distribution, but should be run with a low scan rate to avoid QPS spikes.

Concentrated expiration : Massive keys expiring at the same moment trigger Redis’s active expiration cycle, which runs in the main thread and can add up to 25 ms latency. Searching for EXPIREAT or PEXPIREAT in code and adding a random offset (e.g., redis.expireat(key, expire_time + random(300))) spreads the load.

Memory limit and eviction : When maxmemory is reached, Redis evicts keys according to the configured policy (e.g., allkeys‑lru, volatile‑lru). Evicting large keys further increases latency; splitting data across multiple instances can alleviate pressure.

Fork overhead : RDB/AOF generation and full‑sync create a child process that copies the page table. Large memory footprints cause long fork pauses, observable via INFO field latest_fork_usec. Scheduling backups during off‑peak hours or disabling AOF for non‑critical data reduces impact.

CPU binding : Binding Redis to a specific CPU core makes the forked child compete for the same core, worsening latency during persistence. Avoid CPU pinning when persistence is enabled.

AOF configuration : The three fsync policies ( always, everysec, no) trade durability for performance. everysec is generally recommended as it limits I/O while losing at most one second of data.

Swap usage : If the host swaps, Redis performance collapses. Monitoring memory and swap, and freeing swap by restarting or failover, is essential.

Network saturation : High network I/O can also cause latency spikes. Monitoring NIC throughput and scaling or migrating instances when bandwidth is saturated helps maintain low latency.

In summary, Redis latency can stem from command complexity, large keys, expiration patterns, memory eviction, persistence forks, CPU binding, AOF settings, swap, or network overload; understanding these mechanisms and applying the appropriate mitigations keeps Redis performant.