DAta Feeder For You = DAFFY

Whatisit

DAFFY is a simple data line feeder. It is designed to accept data sets as uploads (either in CSV or JSON format), then serve individual rows up with each GET /next request.

The primary intended use is in large-factor functional stress testing. By using DAFFY, you just feed your generated data set into the server, then your load-testing agents (which may even be distributed) will get a new data row by each subsequent GET /next calls - that is, until the data set runs out.

You can have as many separate data sets in DAFFY as you wish, but bear in mind that DAFFY keeps everything in memory.

Installation

Local

npm i && DAFFY_PORT=<your port> node ./daffy.js

If you omit DAFFY_PORT the default is 8033.

Docker

Use subpardaemon/daffy:latest in your docker-compose file, and if the default 8033 port does not suit you, make sure you set the DAFFY_PORT envvar to a value of your preference.

Usage

Whenever a dataset ID is used, it must match this regex: /^[a-zA-Z0-9_:,-]+$/.

GET /<id>/next or GET /<id>/next/<number>

Get the next row or the next <number> rows in dataset <id>. If you don't specify a number, you get the data row as a JSON-encoded object; if you specify a number (even if it is 1) you receive the data rows as a JSON-encoded array.

GET /list

Get information about current datasets (id, count, repeat and current index). The info is sent in JSON format.

POST /<id> or POST /<id>/<repeats> or PUT /<id>

Upload a dataset under ID <id>. If <repeats> is specified, it means the dataset can be fetched <repeats> times instead of the default, which is each element can be fetched only once. If the request is sent as PUT, it appends the new rows instead of replacing the dataset.

By default, it expects data to be in JSON. If you want to upload CSV, make sure you set the Content-type header to text/csv. Otherwise, just use application/json.

WARNING: both POST and PUT operations happen synchronously, so make sure you're not fetching data from DAFFY while uploading a new dataset.

If the upload was a success, it retuns a small JSON object with statistics data.

GET /

Loads an index page. This will soon be replaced with a crude admin interface to upload data sets and check the current state of affairs within Daffy.

GET /ping

Healthcheck endpoint.

Roadmap

• Human user interface defined in index.html
• Healthcheck in Docker
• Maybe an XML input/output vartiant would be nice. Maybe. Does anyone still use XML out there? Hello?
• Maybe an asynchronous upload?