Skip to content

Deployment

The Publisher software can be deployed in several configurations depending on your operational requirements, compliance needs, and team preferences.

Managed Cloud

K42 hosts and operates a dedicated Publisher instance for you. Each publisher gets a single-tenant deployment - your data is fully isolated.

  • K42 handles provisioning, upgrades, scaling, and availability
  • You receive an endpoint URL to connect your backend services
  • No infrastructure management required on your side

Best for teams that want zero operational overhead and are comfortable with data residing in K42's infrastructure.

Bring Your Own Cloud (BYOC)

The Publisher software is deployed into your own cloud account. You provide K42 with access to a project (GCP) or account (AWS), and K42 deploys and manages the resources within it.

  • Data stays within your cloud boundary
  • K42 manages the deployment, upgrades, and operations
  • You retain full visibility and control over the underlying resources
  • Single-tenant, dedicated infrastructure

Best for organisations with data residency requirements or compliance policies that require data to remain within their own cloud estate, while still wanting managed operations.

Self-Hosted

You deploy and operate the Publisher software yourself on your own infrastructure. K42 provides the software artifact in the format that suits your stack:

  • Docker image - available via private registry credentials or provided via an OCI image
  • JAR file - for JVM-based deployments
  • Native binary - a statically compiled binary (no JVM required)

You are responsible for provisioning, networking, upgrades, and maintenance. K42 provides documentation and support.

Best for organisations with strict compliance requirements, custom infrastructure, or teams that prefer full operational control.

Local Development

For development and testing, run the Publisher software on your own machine. The same artifacts available for self-hosted deployments work locally:

  • Run the Docker image or use a provided Docker Compose example
  • Run the JAR directly on a local JVM
  • Run the native binary