https://github.com/cguertin14/load-tester
Load-tester Docker image.
https://github.com/cguertin14/load-tester
docker load-testing lua
Last synced: 3 months ago
JSON representation
Load-tester Docker image.
- Host: GitHub
- URL: https://github.com/cguertin14/load-tester
- Owner: cguertin14
- License: apache-2.0
- Created: 2021-07-02T04:56:38.000Z (almost 5 years ago)
- Default Branch: main
- Last Pushed: 2022-11-28T21:02:29.000Z (over 3 years ago)
- Last Synced: 2023-03-04T05:04:00.946Z (about 3 years ago)
- Topics: docker, load-testing, lua
- Language: Lua
- Homepage:
- Size: 11.7 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# load-tester
Load-tester Docker image. Fork of `docker.io/argoproj/load-tester` using Alpine Linux.
## Included tools
This directory contains the Dockerfile for `cguertz/load-tester`, which is a simple image that contains:
* [`wrk`](https://github.com/wg/wrk) - an HTTP benchmarking tool
* `report.lua` - a custom wrk reporting script, which generates a report in json format
* [`jq`](https://github.com/stedolan/jq) - JSON processing utility to interpret the report and analyze for success or failure
## Usage
`wrk` can be invoked as follows:
```bash
# Run a benchmark for 45 seconds, using 10 threads, keeping 50 HTTP connections open, and generate a report
$ wrk -t10 -c50 -d45s -s report.lua http://your-http-service/test
Running 45s test @ http://your-http-service/test
10 threads and 50 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 73.28ms 122.80ms 1.21s 89.58%
Req/Sec 171.87 229.29 3.77k 92.52%
55047 requests in 45.28s, 8.24MB read
Non-2xx or 3xx responses: 2895
Requests/sec: 1215.71
Transfer/sec: 186.39KB
```
This generates the following `report.json`:
```json
{
"duration_seconds":45.28,
"requests_total":55047,
"requests_per_second":1215.71,
"transfer_bytes_total":8642337,
"transfer_bytes_per_second":190865.64,
"errors_total":2895,
"errors_ratio":0.05,
"latency_min_ms":0.05,
"latency_max_ms":1206.71,
"latency_avg_ms":73.28,
"latency_stdev_ms":122.80
}
```
The `report.json` file can then be interpreted using `jq`:
```bash
# Ensure error rate was less than 5% and and average latency was under 100ms
$ jq -e '.errors_ratio < 0.05 and .latency_avg_ms < 100' report.json
false
```
The load-tester image can be used as part of a kubernetes job
```yaml
apiVersion: batch/v1
kind: Job
metadata:
name: load-test
spec:
template:
spec:
containers:
- name: load-tester
image: cguertz/load-tester:latest
command: [sh, -c, -x, -e]
args:
- |
wrk -t10 -c40 -d45s -s report.lua http://your-http-service/test
jq -e '.errors_ratio <= 0.05 and .latency_avg_ms < 100' report.json
```