Query, inspect, and manage Redis databases from command line

redis-cli - Redis Command Line Interface documents the primary CLI for Redis, covering command-line execution (run one command and exit), an interactive REPL mode with tab completion and history, and special modes for monitoring, latency analysis, keyspace scanning, and data import/export. A prerequisite check and platform install commands are given for macOS, Ubuntu/Debian, CentOS/RHEL, Alpine, building from source, or running via Docker with no local install at all:

### Check if redis-cli is installed
redis-cli --version

It documents connection options (default host/port, password via -a or a URI, TLS with --cacert, a specific database index via -n, IPv4/IPv6 preference), and a full data-query cheat sheet across every Redis type with Big-O complexity noted per command: string (GET/SET with NX|XX/EX|PX|KEEPTTL, INCR, MGET), hash (HGET, HGETALL, HSCAN), list (LRANGE, LLEN, LPOS), set (SMEMBERS, SISMEMBER, SSCAN), sorted set (ZRANGE by index, score, or lex, ZRANK, ZSCAN), and key inspection (EXISTS, TYPE, TTL, MEMORY USAGE, OBJECT ENCODING, DBSIZE). SCAN-based iteration is documented as the production-safe alternative to KEYS *, with the specific guarantee that a full cursor-0-to-cursor-0 iteration returns every element that existed for the entire scan duration, though elements may repeat and the caller must dedupe.

Server-inspection tooling covers --stat for live per-second stats, the INFO sections for server, memory, keyspace, and replication, --bigkeys/--memkeys/--keystats for keyspace size analysis, and four latency tools - --latency, --latency-history, --latency-dist, and --intrinsic-latency for the host's own baseline. Output can be forced raw, human-readable, CSV, or JSON (RESP3, with -2 forcing RESP2), and a command can be repeated N times (-r) with an optional delay (-i). Server administration commands span ACL management (ACL SETUSER/DELUSER/GENPASS/DRYRUN), client management (CLIENT LIST/KILL/PAUSE/SETNAME), live configuration (CONFIG GET/SET/REWRITE/RESETSTAT), replication acknowledgment (WAIT, and WAITAOF for Redis 7.2+), persistence (BGSAVE, BGREWRITEAOF, LASTSAVE), replica promotion (REPLICAOF), shutdown (SHUTDOWN SAVE/NOSAVE), slow-log inspection, and cluster health and resharding commands.

Use it whenever a task involves querying, inspecting, debugging, or managing Redis from the command line - key/value reads and writes, SCAN or keyspace exploration, server monitoring, or administration. It carries explicit safety guidance: never pass a password via -a in production since it's visible in shell history and process listings, use the REDISCLI_AUTH environment variable instead; never run KEYS * on a large production database since it blocks the server, use SCAN instead; use MONITOR cautiously and never for extended periods on production, since it logs every command including sensitive data; verify the target database with CLIENT LIST or INFO keyspace before FLUSHALL/FLUSHDB, since both are irreversible; and be aware that an --rdb transfer taken during active writes on a busy server can produce an inconsistent snapshot. Seven detailed reference files exist for depth beyond the inline quick reference and are read on demand: references/connection-and-options.md, references/data-query-commands.md, references/module-data-types.md covering RedisJSON, Vector Sets, Bloom and Cuckoo filters, TimeSeries, and RediSearch, references/key-management.md, references/inspection-and-monitoring.md, references/advanced-features.md covering Lua scripting, Pub/Sub, and cluster management, and references/server-administration.md.

Three documented common workflows: exploring an unknown database (INFO keyspace and DBSIZE, then --bigkeys/--memkeys, then sampling keys via --scan | head -20 and checking TYPE/TTL, then reading each key by its actual type); monitoring in real time (--stat -i 2 for live stats, a filtered INFO memory loop, MONITOR with an explicit high-overhead caution, and --latency-history -i 5); and querying specific key patterns, counting keys matching a glob via --scan --pattern '...' | wc -l, or looping over matched keys with a shell while read to inspect each one's hash contents or TTL. Output formatting for scripting includes --raw, the default when piping, --csv, and --json, plus reading the last argument from stdin with -x.

Ships with Redis itself (redis-tools on Debian and Ubuntu) and connects over the standard Redis protocol, with TLS support via --tls --cacert, URI-style connections (redis://user:password@host:port/dbnum), and cluster-aware connections via -c. Its module-data-types reference covers integration with RedisJSON, RediSearch (FT.*), and RedisTimeSeries (TS.*) commands when those modules are loaded on the server.

Anyone operating a Redis instance from the terminal - exploring an unfamiliar database, monitoring live performance and latency, scanning or analyzing the keyspace, or performing server administration like ACLs, configuration, replication, or persistence - who wants the safe, production-appropriate command for each task, such as SCAN over KEYS and REDISCLI_AUTH over -a, rather than re-deriving it each time.