Dasher: Your Agile Feature Toggle and Experimentation Engine

Dasher empowers startups and agile teams to rapidly iterate, test, and validate ideas without costly upfront development.

Key Features

🚩 Feature Flags: Enable/disable functionality at runtime.

Upcoming Features

A/B Testing: Compare multiple versions and track results.
Percentage Rollouts: Gradually increase exposure to new features.
User Targeting: Deliver personalized experiences based on user attributes.
Metrics Tracking: Integrate with your analytics platform for in-depth analysis.
Audit Logs: Maintain a record of changes for transparency.

Why Dasher?

De-risk Innovation: Safely launch new features or UI changes to a subset of users. Gather real-world feedback and data before full rollout.
Data-Driven Decisions: Run A/B experiments to compare variations and measure impact on key metrics. Make informed choices about what works best.
Simple Integration: Lightweight library designed for minimal disruption to your existing codebase. Get up and running quickly.
Flexible Control: Define targeting rules based on user segments, demographics, or even custom criteria.
Real-time Management: Adjust feature flags and experiment parameters on the fly through a user-friendly dashboard (or API).

Who Should Use Dasher?

Product Teams: Experiment with new features and gather data to inform the product roadmap.
Marketing Teams: Test different messaging or promotional strategies to optimize campaigns.
Engineering Teams: Mitigate the risk of releasing new code by gradually rolling it out.

Preview

Quick Start

Using Helm Chart

Look at the chart for more information about possible configuration parameters.

helm repo add dasher https://flohansen.github.io/dasher
helm repo update
helm install <release-name> dasher/dasher

Using Kustomize

By cloning the repository

git clone https://github.com/flohansen/dasher && cd dasher
kubectl apply -k deployment/kubernetes

By using github directly

kubectl apply -k https://github.com/flohansen/dasher/deployment/kubernetes

Using docker

You can use the official Docker image. Because Dasher is a stateful application, make sure to mount a volume to the container. The following example uses the current working directory as data volume.

docker run --name my-dasher-server \
    -p 3000:3000 \
    -p 50051:50051 \
    -v .:/data \
    ghcr.io/flohansen/dasher-server:latest

Use it in your application

go get github.com/flohansen/dasher

package toggles

import (
    "github.com/flohansen/dasher/pkg/dasher"
)

var (
    VerboseLogging = dasher.NewFeature(dasher.FeatureOptions{
        Name: "USE_VERBOSE_LOGGING",
        Description: "If the application should log more verbose",
    })

    // Add more features ...
)

func init() {
    // This registers the feature. The dasher server will create a toggle, if
    // it does not exist.
    dasher.MustRegister(VerboseLogging)

    // Register more features ...
}

package main

import (
    "context"
    "log"
    "os"

    "github.com/flohansen/dasher/pkg/dasher"
)

var (
    dasherServerAddr = os.Getenv("DASHER_SERVER_ADDR")
)

func main() {
    ctx := context.Background()

    // This will subscribe to any changes related to the defined features. On
    // every change the states of the features will be updated.
    go dasher.Listen(ctx, dasherServerAddr)

    for {
        // Check the state of the feature toggle. The "Enabled" property is
        // being synchronized with the feature toggle state of the server.
        if toggles.VerboseLogging.Enabled {
            log.Println("wait one second...")
        }

        time.Sleep(time.Second)
    }
}