bench

README

MQTT Benchmarking Tool

A simple MQTT benchmarking tool for Mainflux platform.

It connects Mainflux things as subscribers over a number of channels and uses other Mainflux things to publish messages and create MQTT load.

Mainflux things used must be pre-provisioned first, and Mainflux provision tool can be used for this purpose.

Installation

cd tools/mqtt-bench
make

Usage

The tool supports multiple concurrent clients, publishers and subscribers configurable message size, etc:

./mqtt-bench --help
Tool for extensive load and benchmarking of MQTT brokers used within Mainflux platform.
Complete documentation is available at https://mainflux.readthedocs.io

Usage:
  mqtt-bench [flags]

Flags:
  -b, --broker string     address for mqtt broker, for secure use tcps and 8883 (default "tcp://localhost:1883")
      --ca string         CA file (default "ca.crt")
  -c, --config string     config file for mqtt-bench (default "config.toml")
  -n, --count int         Number of messages sent per publisher (default 100)
  -f, --format string     Output format: text|json (default "text")
  -h, --help              help for mqtt-bench
  -m, --mainflux string   config file for Mainflux connections (default "connections.toml")
      --mtls              Use mtls for connection
  -p, --pubs int          Number of publishers (default 10)
  -q, --qos int           QoS for published messages, values 0 1 2
      --quiet             Supress messages
  -r, --retain            Retain mqtt messages
  -z, --size int          Size of message payload bytes (default 100)
  -t, --skipTLSVer        Skip tls verification
  -t, --timeout           Timeout mqtt messages (default 10000)

Two output formats supported: human-readable plain text and JSON.

Before use you need a mfconn.toml - a TOML file that describes Mainflux connection data (channels, thingIDs, thingKeys, certs). You can use provision tool (in tools/provision) to create this TOML config file.

Example use and output

Without mtls:

./mqtt-bench --broker tcp://localhost:1883 --count 100 --size 100 --qos 0 --format text --pubs 10 --mainflux mfconn.toml

With mtls ./mqtt-bench --broker tcps://localhost:8883 --count 100 --size 100 --qos 0 --format text --pubs 10 --mainflux mfconn.toml --mtls -ca ca.crt

You can use `config.toml` to create tests with this tool:

./mqtt-bench --config config.toml

Example of `config.toml`:

[mqtt] [mqtt.broker] url = "tcp://localhost:1883"

[mqtt.message] size = 100 format = "text" qos = 2 retain = true

[mqtt.tls] mtls = false skiptlsver = true ca = "ca.crt"

[test] pubs = 3 count = 100

[log] quiet = false

[mainflux] connections_file = "mfconn.toml"

Based on this, a test scenario is provided in `templates/reference.toml` file.

Documentation

Index

• func Benchmark(cfg Config)
• type Client
• type Config
• type JSONResults

Functions

func Benchmark

func Benchmark(cfg Config)

Benchmark - main benchmarking function

Types

type Client

type Client struct {
	ID         string
	BrokerURL  string
	BrokerUser string
	BrokerPass string
	MsgTopic   string
	MsgSize    int
	MsgCount   int
	MsgQoS     byte
	Quiet      bool

	MTLS       bool
	SkipTLSVer bool
	Retain     bool
	CA         []byte
	ClientCert tls.Certificate
	ClientKey  *rsa.PrivateKey
	SendMsg    handler
	// contains filtered or unexported fields
}

Client - represents mqtt client

type Config

type Config struct {
	MQTT mqttConfig   `toml:"mqtt" mapstructure:"mqtt"`
	Test testConfig   `toml:"test" mapstructure:"test"`
	Log  logConfig    `toml:"log" mapstructure:"log"`
	Mf   mainfluxFile `toml:"mainflux" mapstructure:"mainflux"`
}

Config struct holds benchmark configuration

type JSONResults

type JSONResults struct {
	Runs   []*runResults `json:"runs"`
	Totals *totalResults `json:"totals"`
}

JSONResults are used to export results as a JSON document