ops/advisory-ai/Dockerfile builds WebService and Worker images (multi-role via PROJECT/APP_DLL args).
ops/advisory-ai/docker-compose.advisoryai.yaml runs WebService + Worker with shared data volume; ships remote inference toggle envs.
ops/advisory-ai/helm/ provides a minimal chart (web + worker) with storage mounts, optional PVC, and remote inference settings.

Build images

# WebService
docker build -f ops/advisory-ai/Dockerfile \
  -t stellaops-advisoryai-web:dev \
  --build-arg PROJECT=src/AdvisoryAI/StellaOps.AdvisoryAI.WebService/StellaOps.AdvisoryAI.WebService.csproj \
  --build-arg APP_DLL=StellaOps.AdvisoryAI.WebService.dll .

# Worker
docker build -f ops/advisory-ai/Dockerfile \
  -t stellaops-advisoryai-worker:dev \
  --build-arg PROJECT=src/AdvisoryAI/StellaOps.AdvisoryAI.Worker/StellaOps.AdvisoryAI.Worker.csproj \
  --build-arg APP_DLL=StellaOps.AdvisoryAI.Worker.dll .

Local/offline compose

cd ops/advisory-ai
docker compose -f docker-compose.advisoryai.yaml up -d --build

Set ADVISORYAI__INFERENCE__MODE=Remote plus ADVISORYAI__INFERENCE__REMOTE__BASEADDRESS/APIKEY to offload inference.
Default mode is Local (offline-friendly). Queue/cache/output live under /app/data (binds to advisoryai-data volume).

Helm (cluster)

helm upgrade --install advisoryai ops/advisory-ai/helm \
  --set image.repository=stellaops-advisoryai-web \
  --set image.tag=dev \
  --set inference.mode=Local

Enable remote inference: --set inference.mode=Remote --set inference.remote.baseAddress=https://inference.your.domain --set inference.remote.apiKey=<token>.
Enable persistence: --set storage.persistence.enabled=true --set storage.persistence.size=10Gi or --set storage.persistence.existingClaim=<pvc>.
Worker replicas: --set worker.replicas=2 (or --set worker.enabled=false to run WebService only).

Operational notes

Data paths (/app/data/plans, /app/data/queue, /app/data/outputs) are configurable via env and pre-created at startup.
Guardrail phrases or policy knobs can be mounted under /app/etc; point ADVISORYAI__GUARDRAILS__PHRASESLIST to the mounted file.
Observability follows standard ASP.NET JSON logs; add OTEL exporters via OTEL_EXPORTER_OTLP_ENDPOINT env when allowed. Keep disabled in sealed/offline deployments.
For air-gapped clusters, publish built images to your registry and reference via --set image.repository=<registry>/stellaops/advisoryai-web.