groot
e0bf42bf32
mega-ci / static-release-gates (push) Failing after 11s
release-tag / release-image (push) Successful in 6m46s
mega-ci / go-quality (services/control) (push) Successful in 10m7s
mega-ci / go-quality (platform/neuroforge) (push) Successful in 10m20s
mega-ci / go-quality (services/agent) (push) Successful in 11m1s
mega-ci / go-quality (services/knowledge) (push) Successful in 11m13s
mega-ci / docker-build (push) Has been skipped
GPU Subagent + Ollama
Runs Ollama on the GPU node, pulls the configured chat and embedding models through two idempotent init services, and registers a GPU worker with capabilities model.chat and model.embed.
Prerequisites: NVIDIA driver, NVIDIA Container Toolkit and working Docker GPU access.
Edit NEUROFORGE_MASTER_URL, then:
./preflight.sh
docker compose --profile monitoring up -d
Ollama is intentionally exposed on port 11434 so the Master Agent/Knowledge services can use this GPU node directly. Restrict ports 11434, 9100 and 9400 to the Master/Prometheus host with a firewall.