PROJECT 01 / SYSTEMS / 2026

The Home Platform.

Proxmox guests and Docker services, organized by workload.

Architecture sketch / Workload placement and day-to-day platform operations
01 / THE PROBLEM

Problem.

My media services, home automation, and development experiments needed different operating environments. I wanted to change a lab or test a tool without treating the production services as part of the same experiment.

02 / THE APPROACH

Implementation.

  1. 01

    Used Proxmox VMs and containers to give production services, development work, media serving, and lab experiments distinct environments. Workload placement follows the operating needs of each service rather than treating every guest the same.

  2. 02

    Managed containerized production services through Docker Compose, while keeping workloads with different storage or device needs in their own guests. Defined storage bindings deliberately instead of assuming guest disks and shared data are interchangeable.

  3. 03

    Kept administration separate from the sandbox workflow. Shared project knowledge helps tools understand the environment; it does not replace an explicit decision about their execution authority.

  4. 04

    Tracked configuration changes, workload checks, and deferred migration steps in operational notes. Validated services at the guest and host layers so an infrastructure problem is not mistaken for an application failure.

Virtualization layer
Workload environments
Service lifecycle
03 / THE RESULT

Result.

Established distinct production and experimental environments, with separately managed media and smart-home workloads and a dedicated administration workflow. Subsequent lab provisioning and service integrations build on that platform rather than sharing one undifferentiated host.

VERIFICATION

Documented guest placement, Docker service checks, dedicated workload migrations, and later lab builds. The illustration summarizes workload roles; operational inventory and access configuration remain private.

04 / THE TAKEAWAY

What I learned.

A VM boundary is one part of an operating model. Network policy, credentials, storage access, and human approval each need their own treatment. Useful documentation explains those decisions while keeping the live access map private.

Questions about this project?

Email Binh ↗
NEXT PROJECT

Project Hestia