Admin Documentation

Admin documentation is for people who install and operate a prokube deployment, connect it to organizational infrastructure, or expose custom workloads through the platform. User-facing concepts that apply across Labs, MLOps, and AgentOps are documented under Foundation.

What Admins Own

• deployment prerequisites such as Kubernetes, DNS, TLS, storage, identity, and registry access;
• workspace, user, group, and network-policy administration;
• external identity-provider and custom application integration;
• capacity-sensitive resources such as storage, GPUs, databases, and file storage;
• backup, restore, upgrade, observability, and incident runbooks.

Available Pages

Page	Use it for
Installation Planning	Planning Kubernetes, DNS, TLS, identity, registry, and storage prerequisites before installation.
User Management	Managing users, workspaces, groups, workspace access, and workspace security policy settings.
Identity Providers	Connecting external SSO providers to Keycloak and troubleshooting login/group mapping.
Network Policies	Defining reusable egress profiles and assigning outbound network restrictions to workspaces.
Application Networking	Exposing custom applications through prokube gateways and Istio routing.
Application Authentication	Choosing gateway authentication or direct OIDC for custom applications.
External Models	Connecting centrally managed model providers and granting individual models to workspaces.
Storage	Understanding StorageClasses, local storage, replicated storage, and PVC troubleshooting.
GPU Administration	Operating GPU nodes, NVIDIA GPU Operator, timeslicing, MIG, and GPU monitoring.
Backup and Restore	Disaster-recovery scope, backup storage, restore expectations, and validation.
Upgrading	Planning release upgrades, GitOps-driven rollout, and post-upgrade validation.
Operations Runbooks	Administrator runbooks for MinIO, Keycloak bootstrap, and observability operations.
System Status	Checking administrator-only backend component health and backend metadata; the same page also shows workspace pod quota to all users.

Platform-wide monitoring is an administrator topic. Use System Status for quick backend health checks. For dashboards, alerts, and log retention, use the deployment's Grafana, Prometheus, Alertmanager, and Loki.