26 lines
1.1 KiB
Markdown
26 lines
1.1 KiB
Markdown
## Benchmark Latency and Throughput of Stitcher
|
|
### Benchmark Latency
|
|
0. Have Python installed and the deps : pandas and matplotlib
|
|
1. Go to `cd latency`
|
|
2. Run this in the terminal
|
|
```
|
|
./request.sh -o <output_location> -u "<backend_url>" -i <iteration> --no-output
|
|
```
|
|
- `-o` specifies the output folder
|
|
- `-u` backend url e.g: `http://stitchaton.local`
|
|
- `-i` (optional) specifies the range of the iteration e.g: 1-11 or just 3
|
|
- `--no-output` dumb the output to /dev/null
|
|
|
|
|
|
### Benchmark Throughput
|
|
0. Have JMetter installed, [download here](https://jmeter.apache.org/download_jmeter.cgi)
|
|
1. Go to `cd throughput`
|
|
2. Run this in the terminal
|
|
```
|
|
./request.sh -o <output_location> -H "<backend_ip>" -t <num_of_threads> -l <num_of_loops> -p <optional_port_default5000>
|
|
```
|
|
- `-o` specifies the output folder
|
|
- `-H` backend api without http or port e.g: 10.250.123.123 or stitchapi.local
|
|
- `-t` number of concurrent requests or threads
|
|
- `-l` number of loops of each concurrent requests e.g: 10 threads and 20 loops will result in 200 requests in total
|
|
- `-p` port number in which the application is served
|