Overview

The project aims to bring the feature of runtime configuration to the masses. The configuration is list of options helping you to control the behaviour of your applications. More precisely configset provides following features:

1. Manages the configuration in the single dedicated place.
2. Distributes configuration across different physical servers/VMs/containers using language-specific SDK. Holds configuration state in memory.
3. Allows people from of a team (DevOps, managers, developers) to observe and manage the configuration using a web dashboard.
4. Tracks changes of the configuration and provides "listening" API to call application code to reflect the change ( update internal state, call a function, etc.).

A structure of the configuration is straightforward: the configuration is split by applications. Each application has many properties. Each property is a list of (hostname, key, value) triples. Each client has hostname parameter. When a client asks for a particular property in a config application then only value for a given hostname is returned (or default one if configured).

There is no difference to the actual purpose of configuration: infrastructure, application features, secrets, etc.

Architecture

There are four parts: a client library, a server application and a dashboard.

1. Client connects to gRPC server application, receives a current configuration snapshot, listens for future configuration changes.
2. Server stores configuration in some sort of storage (by now two options exist: in memory for tests or external PostgreSQL) and handles outgoing streams to connected clients. At every point of time it knows which client is looking for which configuration.
3. CRUD-like UI to apply changes to the system.

Usage example

Sample project

To run the sample (docker-compose is required) execute locally ./sample_run.sh. Dashboard is available at http://localhost:8080/

Development

1. ./local_run.sh - starts db/dashboard/server in docker-compose locally. A dashboard will be available at http://localhost:8080/.
2. ./gradlew test - runs tests across the project.

Deployment

1. Build images - ./gradlew server:dockerBuildImage && ./dashboard:dockerBuildImage. And push them into the registry.
2. Deploy (or use existed) one of the available storage backend. Buy now Postgresql is the only option. PRs are welcome!
3. Deploy server and dashboard to your infrastructure using compose file as an example.
4. Publish Java SDK to your maven repository, add artifact to your project.

License

Apache-2.0