Docker Services
A dedicated Docker environment used to host AI infrastructure, application services, uptime monitoring, supporting data services, and retained observability tooling within the home lab.
Project Goal
The goal of this environment is to maintain a dedicated Docker platform for AI services, application workloads, supporting data services, monitoring, and infrastructure tooling without mixing unrelated workloads directly into other lab systems.
Public documentation is intentionally sanitized. Internal IP addresses, exact port mappings, private Compose configuration, credentials, administrative URLs, and provider authentication details are not published.
Environment Summary
Container Platform
Docker 29.8.1 running inside a dedicated Debian 13 LXC container hosted on Proxmox VE.
Docker Compose
Docker Compose v5.5.1 is used to organize AI, application, uptime-monitoring, and observability workloads as separate projects.
OmniRoute AI Gateway
OmniRoute runs as a healthy self-hosted AI gateway/router with a dedicated Redis backend, providing a controlled routing layer between client tools and external AI model services.
AI Provider Integration
End-to-end model routing through OmniRoute has been successfully validated with NVIDIA NIM while keeping provider authentication and internal configuration private.
Uptime Monitoring
Uptime Kuma currently runs as a healthy service for availability monitoring and service-status visibility.
Retained Observability Stack
Prometheus, Grafana, Alertmanager, cAdvisor, Node Exporter, and the Proxmox exporter remain retained as prior observability work but are currently stopped.
Architecture
Technologies Used
What I Built
- Deployed a dedicated Debian LXC container for Docker-based application and infrastructure services.
- Installed and validated Docker and Docker Compose for repeatable container deployment and management.
- Deploy and operate OmniRoute as a self-hosted AI gateway/router with a dedicated Redis backend using Docker Compose.
- Successfully validated end-to-end AI model routing through NVIDIA NIM using OmniRoute.
- Deploy and currently operate Uptime Kuma for availability monitoring and service-status visibility.
- Built and retained an observability stack using Prometheus, Grafana, Alertmanager, cAdvisor, Node Exporter, and a Proxmox exporter.
- Organized workloads into separate Compose-managed projects for AI routing, supporting services, uptime monitoring, and observability.
- Use Docker networking and health checks to manage service separation and validate active workloads.
- OmniRoute, Redis, and Uptime Kuma currently report healthy Docker states.
Security and Privacy Considerations
- Exact internal addressing, port mappings, private URLs, credentials, and Compose file contents are not published.
- AI provider credentials and authentication details remain private.
- The Docker services environment is separated from other lab workloads to simplify administration and reduce unnecessary operational overlap.
- Administrative and monitoring interfaces are treated as internal infrastructure services rather than general public services.
- Public documentation focuses on architecture, service design, lifecycle management, and operational value without exposing sensitive configuration.
Operational Value
This environment demonstrates day-to-day Docker operations across AI gateway hosting, application services, supporting data services, uptime monitoring, Compose project organization, health-state validation, Docker networking, and service lifecycle management.
OmniRoute adds a practical AI infrastructure component by providing a self-hosted routing layer that can be tested and managed independently of individual client tools or model services.
The retained observability stack also documents prior hands-on work with Prometheus, Grafana, Alertmanager, cAdvisor, Node Exporter, and Proxmox metrics collection without implying that those services are currently online.
What This Demonstrates
The project demonstrates containerized AI infrastructure, Docker Compose operations, AI provider routing, Redis-backed services, uptime monitoring, Docker networking, container health checks, service lifecycle management, and prior observability-stack deployment.
Project Status
Active and operational. OmniRoute, Redis, and Uptime Kuma are currently running and report healthy. The Prometheus/Grafana observability stack and Proxmox exporter remain retained but stopped and would be revalidated before being returned to service.