EEM-Hub/openshift-expert

OpenShift Expert

Expert knowledge base for OpenShift Container Platform architecture, operations, and deployment patterns. Contains 3,820 justified beliefs covering the operator-driven platform model, networking, virtualization, security, edge deployments, and 50+ subsystems.

What is this?

This is an External Epistemic Memory (EEM) — a model-agnostic knowledge base that any LLM can use via the reasons CLI or tool calling. Unlike a LoRA or fine-tune, this knowledge is not baked into model weights. It is external, inspectable, correctable, and works with any model.

Stats

Metric	Value
Total beliefs	3,820
Status	3,820 IN / 0 OUT
Premises (observations)	3,670
Derived (justified conclusions)	150
Nogoods (contradictions)	0
Retraction rate	0%
Max derivation depth	7

Top Topics

Topic	Beliefs
ocp	560
default	312
api	224
cluster	192
operator	192
openshift	159
namespace	124
node	120
hcp	110
config	108
network	99
image	91
virt	91
install	85
supported	74
policy	72

Domain Coverage

• OCP Core Platform: operator-driven immutable model, cluster versioning, install-time constraints, platform topology variants (511 beliefs)
• Hosted Control Planes (HCP): multi-tenant control planes, NodePool management, hosted cluster lifecycle (107 beliefs)
• CLI & API: oc commands, API resources, resource management, debugging (63 beliefs)
• OpenShift Virtualization (CNV): VM lifecycle, live migration, virtual networking, storage for VMs (82 beliefs)
• Networking — OVN-Kubernetes: default CNI, pod networking, service routing, hybrid overlay (27 beliefs)
• Networking — Policy & Firewalls: network policies, egress firewalls, admin network policies, user-defined networks (40 beliefs)
• Networking — SR-IOV: hardware-accelerated networking, VF management, node policies (28 beliefs)
• Networking — Ingress & Routes: HAProxy ingress controller, route configuration, TLS termination (34 beliefs)
• Networking — IPsec & Encryption: IPsec tunnel configuration, node-to-node encryption (12 beliefs)
• Operators & OLM: Operator Lifecycle Manager, OperatorHub, File-Based Catalogs, bundle management (87 beliefs)
• Etcd: cluster database, backup/restore, performance tuning, defragmentation (37 beliefs)
• Security — SCCs & RBAC: Security Context Constraints, role-based access control, project isolation (28 beliefs)
• Security — OAuth & Authentication: OAuth server, identity providers, token management (24 beliefs)
• Security — Webhook Admission: validating/mutating webhooks, enforcement model (16 beliefs)
• Builds & Images: S2I builds, Shipwright, image streams, image registry, samples operator (70 beliefs)
• CI/CD & GitOps: Jenkins pipelines, OpenShift GitOps (ArgoCD), deployment strategies (36 beliefs)
• Edge — ZTP & SNO: Zero Touch Provisioning, Single Node OpenShift, TALM upgrades, vDU workloads (53 beliefs)
• Node Management: MachineConfig Operator, machine sets, autoscaling, health checks, tuned profiles (57 beliefs)
• Observability: monitoring stack, logging, Network Observability (NetObserv), flow collection (54 beliefs)
• Console: web console, console plugins, developer/admin perspectives (43 beliefs)
• Storage: CSI drivers, dynamic provisioning, persistent volumes (10 beliefs)
• Backup & DR: OADP, etcd backup, cluster restore (13 beliefs)
• Platform Variants: Nutanix, IBM Cloud, IBM Power, bare metal, ROSA, OCI (101 beliefs)
• RHCOS & Node OS: Red Hat CoreOS, immutable OS, ignition, ostree updates (13 beliefs)
• Additional topics: NMState, MetalLB, templates, jobs, probes, ConfigMaps, HPA, PDB, scheduler, RHOSO, RHOAI (remaining beliefs)

How to Use

Import into a reasons database

reasons init
reasons import-json network.json

Query beliefs

reasons search "operator lifecycle"
reasons explain operator-driven-immutable-platform-model
reasons show install-time-irreversible-constraints

Use as an MCP tool or CLI

Any LLM agent that can call reasons search, reasons show, and reasons explain can use this knowledge base. The agent does not need to be told it is an expert — the knowledge base speaks for itself.

Key Beliefs

Node	Summary
operator-driven-immutable-platform-model	The entire OpenShift platform operates through an operator-driven model with immutable infrastructure
install-time-irreversible-constraints	Multiple cluster-defining decisions are permanently locked at install time and cannot be changed
observability-follows-platform-enablement-pattern	OpenShift observability is a specific instance of the platform-wide multi-component enablement pattern
edge-fleet-management-pipeline	Edge cluster lifecycle follows a managed pipeline: ZTP + GitOps provisions, TALM orchestrates upgrades
multi-cni-network-architecture	OpenShift uses a layered CNI architecture: OVN-Kubernetes primary, Multus for additional networks
webhook-admission-enforcement-model	Webhook admission follows a constrained enforcement model with TLS requirements
node-config-immutable-delivery-pipeline	Node configuration follows an immutable delivery pipeline via MachineConfig Operator
complete-software-delivery-from-build-to-console	OpenShift provides a unified software delivery model covering application and operator lifecycles
platform-model-with-topology-variants	Single operator-driven platform model recognizing two sanctioned topology variants
consoleplugin-compat-level-1	ConsolePlugin has Compatibility Level 1: stable for at least 12 months or 3 minor releases

Sources

Built from exploration of OpenShift Container Platform documentation, API references, and operational experience across networking, virtualization, security, edge, and platform management subsystems.

Files

File	Description
network.json	Full belief network (machine-readable, portable)
reasons.db	SQLite database (gitignored, regenerate with reasons import-json network.json)
CLAUDE.md	Agent instructions for using this knowledge base
entries/	904 exploration entries — raw observations behind the premises

Quality

• All 3,820 beliefs are IN (none retracted)
• 3,670 premises grounded in direct observations of OpenShift behavior and documentation
• 150 derived beliefs justified from premises via SL justifications
• 0 nogoods — no contradictions detected
• Max derivation depth of 7, indicating multi-step reasoning chains
• Built and reviewed using ftl-reasons derive and review-beliefs pipeline

Limitations

• Focused on OpenShift Container Platform as of mid-2026
• OpenShift evolves across minor releases; some beliefs may become stale
• Heavier coverage of networking, virtualization, and operators than storage or serverless
• Platform variant coverage (Nutanix, IBM, bare metal) varies in depth
• No ATMS or assumption-based beliefs (single-context TMS only)

Authors

• Ben Thomasson (@benthomasson)

License

mit