Operations Center

Architecture overview.

Self-hosted. Secure. Extensible. Built on open-source components and designed for enterprise security boundaries and scale.

Operations Center architecture A user request enters the request API, which drives LLM orchestration and agent services. Those services call automation content, testing, security and compliance services. Results flow to Git and work management, and all evidence is stored in the data layer. Monitoring and logging observe every layer. User / automation request API & orchestration layer FastAPI LLM orchestration & agent services PIPELINE SERVICES Automation content Quality & lint Testing & runtime Security Compliance Git & work management — pull request opened Evidence & data layer — artifacts, logs, records Monitoring & logging across every layer
A request enters the API and orchestration layer, which drives LLM orchestration and agent services. Those call the pipeline services — automation content, quality, testing and runtime, security, compliance — and the validated result is delivered as a pull request in Git. All evidence persists in your own data layer, with monitoring and logging across every layer.

TODO — verify the component list below against the product repository before publication. Components not yet released must be labelled as planned.

Layers and responsibilities.

User & request layer

Where a plain-language or structured automation request is submitted and scoped.

Web UI · structured request intake

API & orchestration

Validates the request, owns pipeline state and coordinates every downstream service.

FastAPI

Agent & workflow services

Drives the reasoning steps and the workflow between stages, under explicit tool scopes.

LLM orchestration · n8n

Automation content services

Searches for reusable content, then generates blueprints and Ansible content.

Ansible · roles · collections · Execution Environments

Testing & runtime

Runs linting, Molecule scenarios in disposable environments and live AWX jobs.

Molecule · ansible-lint · AWX · Kubernetes / k3d

Security & compliance

Secret scanning, risky module checks, secure defaults, policy and governance results.

Secret detection · policy checks · secret-manager integration

Git & work management

Opens the pull request, links the work item and carries the evidence trail.

Forgejo · Plane

Evidence & data layer

Stores artifacts, logs, test reports, scan results and change records for audit.

PostgreSQL · MinIO

Monitoring & logging

Pipeline runs, stage outcomes and service health, observable across all layers.

Platform-native observability

Design properties.

Self-hosted

Runs inside your network. Automation content and evidence stay in your boundary.

Secure

Vault-backed secrets, least-privilege tool access, no credentials in generated source.

Extensible

Add your own checks, policies, integrations and Execution Environments.

Audit-ready

Every stage emits retained artifacts tied to the change that produced them.

Human-governed

A pull request is the only output. Merging stays a human decision.

Self-hosted deployment.

Operations Center is deployed into your environment and wired to your own Git, AWX, secret store and identity provider.

  • Kubernetes or k3d deployment
  • Your Git platform for pull-request delivery
  • Your AWX or Ansible Automation Platform for runtime validation
  • Your Vault or secret manager
  • Your identity provider for access control
  • No outbound transfer of automation content

Review the architecture with your platform team.

Bring your security and platform engineers. We will walk the layers, the data boundary and the governance model against your requirements.