redis-benchmark

Run default benchmark

$ redis-benchmark

$ redis-benchmark -h [localhost] -p [6379]

$ redis-benchmark -c [100]

$ redis-benchmark -n [10000]

$ redis-benchmark -t [get,set]

$ redis-benchmark -d [256]

$ redis-benchmark -P [16]

$ redis-benchmark -q

redis-benchmark [-h host] [-p port] [-c clients] [-n requests] [options]

redis-benchmark measures Redis server performance by simulating multiple clients concurrently executing various commands. By default it tests a standard set of operations including SET, GET, INCR, LPUSH, LPOP, and others, reporting requests per second for each command type.
The -c flag controls the number of concurrent client connections, while -n sets the total request count. Pipelining with -P batches multiple requests per round-trip, testing raw throughput independent of network latency. The -d flag adjusts the data payload size to benchmark different workload patterns, and -t limits testing to specific commands.

-h HOST

Server hostname.

Server port.

Concurrent clients.

Total requests.

Test commands.

Data size in bytes.

Pipeline N requests.

Quiet, show query/sec.

Authentication password.

CSV output.

Results vary by hardware and network. Local testing eliminates network overhead. Production has different patterns.

redis-benchmark is included with Redis, originally created by Salvatore Sanfilippo in 2009. It provides standardized performance testing for Redis deployments.

redis-cli(1), redis-server(1), memtier_benchmark(1)