https://github.com/axent-pl/httpster
Simple tool to run HTTP performance tests against multiple endpoints
https://github.com/axent-pl/httpster
ab golang httptrace load-testing wrk
Last synced: 8 months ago
JSON representation
Simple tool to run HTTP performance tests against multiple endpoints
- Host: GitHub
- URL: https://github.com/axent-pl/httpster
- Owner: axent-pl
- License: mit
- Created: 2024-02-11T08:54:59.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-03-01T15:22:43.000Z (over 1 year ago)
- Last Synced: 2025-01-03T15:14:37.856Z (9 months ago)
- Topics: ab, golang, httptrace, load-testing, wrk
- Language: Jupyter Notebook
- Homepage:
- Size: 1.42 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: LICENSE
Awesome Lists containing this project
README
# HTTPster
A simple load test tool that executes multiple requests with any method (GET, POST, PUT, ...) and returns raw data to perform detailed analysis.
## Motivation
I attempted to conduct load tests to evaluate the latency introduced by some selected API Gateways.
Unfortunately, tools like `ab` and `wrk` did not provide sufficient information — offering basically mean and standard error — and their output was less than ideal to work with - formated text.
This prompted the idea to develop a simple tool that could provide raw data for more thorough analysis.
## Usage
```shell
# build
make build
# check for options
bin/httpster --help
# run
cat example.json | bin/httpster -duration=30s -threads=4 1> out/data.csv
```
Please note the `1>` as all the metrics are sent to the `stdout` whereas all the errors to `stderr`.
## Analysis
An example notebook containing analysis of the `HTTPster` output can be found [here](https://github.com/axent-pl/httpster/blob/main/analyse.ipynb).
To load the data from `HTTPster` please start with the following Python snippet:
```python
import pandas as pd
df = pd.read_csv("out/data.csv")
df['StartTime'] = pd.to_datetime(df['StartTime'])
```
From this point you can start your analysis - plot all the distributions, check for anomalies, etc.
## Input format
Request definitions are supposed to be provided via `stdin` in a `JSON` format with following schema:
```json
{
"$schema": "http://json-schema.org/draft-04/schema#",
"type": "array",
"items": [
{
"type": "object",
"properties": {
"id": {
"type": "string"
},
"url": {
"type": "string"
},
"method": {
"type": "string"
},
"headers": {
"type": "object"
},
"body": {
"type": "string"
}
},
"required": [
"url",
"method"
]
}
]
}
```
Example
```json
[
{
"id": "PUT /anything",
"url": "http://httpbin/anything",
"method": "PUT",
"headers": {
"Content-Type": "application/json"
},
"body": "{\"data\": {\"key1\":\"val1\"}}"
}
]
```
## Output format
`HTTPster` ouutput is in `CSV` format to `stdout`. Any error encountered are outputed to `stderr`. Below is an example of the output
```csv
ID,StartTime,Duration_ns,ConnDuration_ns,DialDuration_ns,DNSDuration_ns,Status,StatusCode,Error
GET /,2024-02-12T12:50:58+01:00,338865000,187818875,146996875,40262209,200 OK,200,
GET /,2024-02-12T12:50:58+01:00,342826542,188442916,147601250,40228166,200 OK,200,
GET /,2024-02-12T12:50:58+01:00,343047959,187695292,146739750,40384542,200 OK,200,
GET /,2024-02-12T12:50:58+01:00,484274459,333713000,292769375,40259625,200 OK,200,
GET /anything,2024-02-12T12:50:58+01:00,296632375,147367458,145465667,1757708,200 OK,200,
GET /anything,2024-02-12T12:50:58+01:00,292864292,146642000,145832292,718333,200 OK,200,
GET /anything,2024-02-12T12:50:58+01:00,293629125,146643958,145827709,703833,200 OK,200,
GET /anything,2024-02-12T12:50:58+01:00,299385209,149145125,146590250,2280750,200 OK,200,
PUT /anything,2024-02-12T12:50:59+01:00,298021375,149354792,147107584,2104166,200 OK,200,
PUT /anything,2024-02-12T12:50:59+01:00,297629958,149104042,147283459,1743791,200 OK,200,
PUT /anything,2024-02-12T12:50:59+01:00,296908417,148088459,146979500,987750,200 OK,200,
```