Deterministic, multi-agent AI framework for automated MLOps risk diagnosis and self-healing.
217
Agentic-Ops is an autonomous, deterministic Multi-Agent System (MAS) designed to investigate, diagnose, and remediate operational MLOps signals (Data Drift, Concept Drift, Model Staleness, OOMKilled Pods, Latency Spikes) under strict Site Reliability Engineering (SRE) governance.
Unlike open-ended agentic frameworks, Agentic-Ops enforces a Zero-Trust, Fail-Closed Security Model:
DENY), rate-limiting, and human-in-the-loop (HITL) approval gateways.DataSpecialist, ModelSpecialist, InfraSpecialist, RemediationAgent, WebChatAssistant).The official container image is multi-architecture (linux/amd64 and linux/arm64):
# 1. Pull the official image
docker pull nolll78/agentic-ops:latest
# 2. Run the complete Control Plane
docker run -d -p 8000:8000 --name agentic-ops nolll78/agentic-ops:latest
Once started, access the built-in services in your browser:
You can pass environment variables to configure external LLMs or admin credentials:
docker run -d -p 8000:8000 \
-e ADMIN_API_KEY="your-secret-key" \
-e OPENROUTER_API_KEY="sk-or-v1-..." \
-e DEFAULT_LLM_MODEL="google/gemini-2.0-flash-exp:free" \
--name agentic-ops \
nolll78/agentic-ops:latest
Content type
Image
Digest
sha256:335f5a51d…
Size
211.6 MB
Last updated
about 6 hours ago
docker pull nolll78/agentic-ops