Crates.io | rench |
lib.rs | rench |
version | 0.3.0 |
source | src |
created_at | 2018-02-09 05:08:49.271182 |
updated_at | 2018-06-01 16:26:04.771445 |
description | A simple http benchmark utility. Written in rust... rust + bench = rench |
homepage | https://github.com/kbacha/rench |
repository | https://github.com/kbacha/rench |
max_upload_size | |
id | 50295 |
size | 41,901 |
A benchmark utility for end points. Written in rust... rust + bench = rench
Installation requires either git or cargo.
Clone or fork the repo and run the following:
cargo install -f
cargo install -f rench
This should install and download from the latest release version.
The gist of a http benchmarker is to run a series of queries against an endpoint and collect the facts from it. These facts then are summarized for the user to examine. To really maximize an end point, we use a simple threading model to make many requests at the same time. This allows us to generate a large number of requests and try to fully saturate the http end point.
You can change the number of threads and the number of requests to suit your needs. You can even specify multiple URLs and it will round-robin the requests between them.
$ ./target/release/rench -c 4 --engine=hyper -n 10000 http://0.0.0.0:6767
Beginning requests
1000 requests
2000 requests
3000 requests
4000 requests
5000 requests
6000 requests
7000 requests
8000 requests
9000 requests
10000 requests
Finished!
Took 4.813263821 seconds
2077.5923306697964 requests / second
Summary
Average: 1.88827 ms (std: 0.763338 ms)
Median: 1.012478 ms
Longest: 1005.942426 ms
Shortest: 0.403114 ms
Requests: 10000
Data: 234.38 KB
Status codes:
200: 10000
Latency Percentiles (2% of requests per bar):
▌ 7.399224
▌
▌
▌▌
▖▌▌
▖▌▌▌
▖▌▌▌▌
▖▖▖▌▌▌▌▌▌▌
▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▖▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌
▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌▌ 0
Latency Histogram (each bar is 2% of max latency)
▌ 9862
▌
▌
▌
▌
▌
▌
▌
▌
▌▖ 0
$ rench --help
Git Release Names
Kevin Choubacha <chewbacha@gmail.com>
USAGE:
rench [FLAGS] [OPTIONS] <URL>...
FLAGS:
-h, --help Prints help information
-i, --head The issue head requests instead of get
-V, --version Prints version information
OPTIONS:
--chart-size <chart-size> The size of the chart to render [values: none, n, small, s, medium, m, large, l]
-c <concurrency> The number of concurrent requests to make
-e, --engine <engine> The engine to use [values: hyper, reqwest]
-n <requests> The number of requests in total to make
ARGS:
<URL>... Each url specified will be round robined.