Reference

Technical specifications, APIs, and comprehensive details of all Juju components.

Platform

Juju is an orchestration platform that includes the CLI client, agent binaries, and supporting infrastructure for managing cloud deployments.

Cloud and Charmhub

In Juju, deployments draw from two external sources. Clouds provide compute resources for your infrastructure. Charmhub provides charms – operators for deploying and managing applications.

Client

In Juju, you interact with these resources through clients – command-line and web interfaces for managing controllers, models, and deployments.

Controller

Clients connect to a controller – the central management service that coordinates between clouds, Charmhub, and your deployed resources.

Users

Controller access requires user authentication. User accounts provide authentication and authorization for managing Juju resources.

Infrastructure and applications

Once authenticated, users work with deployments. Within a controller, deployments are organized into models – logical containers for applications, infrastructure, and their supporting components. Each model draws resources from a single cloud.

Models contain applications deployed from charms and composed of units (individual instances).

Applications connect to each other through relations endpoints – integration points between compatible interfaces. Offers enable cross-model relations.

Applications are managed through configuration, secrets, actions, and scripts, and support scaling and high availability. Charms may require resources.

Supporting infrastructure – machines, storage volumes, network spaces and subnets, and availability zones – is provisioned from the cloud. Constraints and placement directives control how resources are selected and allocated. SSH keys provide access.

On each machine, agents (jujud on machines, containeragent on Kubernetes) execute charm code through hooks. Charms use hook commands (provided by jujuc) to interact with Juju. On Kubernetes, containeragent also orchestrates workload containers using Pebble.

Removing and upgrading are cross-cutting operations that apply across multiple resource types – from individual units and applications to entire models and controllers.