Skip to content

02_profiles.md

Latest commit

 

History

History
119 lines (83 loc) · 3.92 KB

02_profiles.md

File metadata and controls

119 lines (83 loc) · 3.92 KB

AIXCL Profiles

Profiles define which operational services are enabled alongside the always-present runtime core.

1. Runtime Core (Always Enabled)

The runtime core consists of two components:

  • Inference Engine (e.g., Ollama, vLLM, llama.cpp) - LLM inference engine (Docker-managed service)
  • OpenCode - AI-powered code assistance (VS Code plugin, not Docker-managed)

OpenCode is a client-side IDE plugin that connects to the Inference Engine via the OpenAI-compatible API. Because it runs inside VS Code rather than as a Docker container, it is not included in the RUNTIME_CORE_SERVICES array or PROFILE_SERVICES mappings in lib/cli/profile.sh.

Note: Runtime persistence is provided by PostgreSQL as an operational service in all current profiles.

2. Profile Overview

Profile Purpose Audience
usr User-oriented runtime End users, minimal deployments
dev Developer workstation Local development
ops Operations-focused Servers/operators
sys System-oriented Complete deployments

3. Profile Definitions

usr

Purpose: User-oriented runtime with minimal footprint, optimized for end-user deployments.

Includes:

  • Runtime core: Inference Engine, OpenCode (plugin)
  • PostgreSQL (database for runtime persistence)

Excludes:

  • Open WebUI (web interface)
  • pgAdmin (database admin)
  • Prometheus, Grafana, Loki, Alloy (monitoring/logging)
  • cAdvisor, node-exporter, postgres-exporter, nvidia-gpu-exporter (metrics)

Use Cases: End-user deployments, resource-constrained environments, minimal installations with database persistence

dev

Purpose: Developer workstation with UI and database tools.

Includes:

  • Runtime core: Inference Engine, OpenCode (plugin)
  • Open WebUI (web interface for model interaction)
  • PostgreSQL (database for conversations and data)
  • pgAdmin (database administration UI)

Excludes:

  • Prometheus, Grafana, Loki, Alloy (monitoring/logging)
  • cAdvisor, node-exporter, postgres-exporter, nvidia-gpu-exporter (metrics)

Use Cases: Local development, testing, interactive model exploration

ops

Purpose: Observability-focused deployment for servers and operators.

Includes:

  • Runtime core: Inference Engine, OpenCode (plugin)
  • PostgreSQL (database for runtime data)
  • Prometheus (metrics collection)
  • Grafana (metrics visualization and dashboards)
  • Loki (log aggregation)
  • Alloy (log shipping)
  • cAdvisor (container metrics)
  • node-exporter (host metrics)
  • postgres-exporter (database metrics)
  • nvidia-gpu-exporter (GPU metrics, if GPU available)

Excludes:

  • Open WebUI (web interface)
  • pgAdmin (database admin)

Use Cases: Production servers, monitoring-focused deployments, observability analysis

sys

Purpose: System-oriented deployment with complete feature set and automation.

Includes:

  • Runtime core: Inference Engine, OpenCode (plugin)
  • All dev services: Open WebUI, PostgreSQL, pgAdmin
  • All ops services: Prometheus, Grafana, Loki, Alloy, cAdvisor, node-exporter, postgres-exporter, nvidia-gpu-exporter

Use Cases: System deployments, demonstrations, full-featured environments with automation

4. Profile Selection Guidelines

  • usr: Use for user-oriented deployments with minimal footprint (runtime core + PostgreSQL)
  • dev: Use for local development and interactive work
  • ops: Use for production deployments requiring observability
  • sys: Use for system-oriented deployments with complete feature set and automatic updates

5. Profile Invariants

All profiles must:

  • Include the complete runtime core (Inference Engine, OpenCode)
  • Never disable or conditionally exclude runtime core components
  • Maintain runtime core independence from operational services

Profiles may:

  • Add operational services as needed
  • Configure operational services differently
  • Exclude operational services based on use case