Zyvor Fabric Documentation Index
Complete documentation index for the Zyvor Fabric infrastructure control plane (Zyvor Fabric daemon).
Getting Started
Guides for new users to install, configure, and begin using Zyvor Fabric.
| Document | Description |
|---|---|
| Getting Started Overview | Section overview and reading order |
| Installation Guide | System requirements, packages, building from source |
| Quick Start | Create your first VM in 5 minutes |
| Configuration Reference | Config file, environment variables, all sections |
| Web UI Guide | Dashboard access, login, VM management through browser |
Tutorials
Step-by-step walkthroughs for common workflows.
| Document | Description |
|---|---|
| Tutorial index | Full numbered tutorial list |
| 01 First VM | Create, configure, start, and connect to a virtual machine |
| 02 Networking | Bridges, VLANs, bonds, port forwarding, network policies |
| 03 Snapshots & Backups | Snapshots, backup policies, restore |
| 04 Advanced VM options | Hotplug, start options, cloud-init |
| 05 Clustering | Datacenters, pools, migration, HA (see also proven-infra limits) |
| 06 Security hardening | Auth, RBAC, TLS, audit |
| 07 Logging & compliance | Journals, scans, secrets |
| 08 OpenStack Clients | Drive Fabric with openstack CLI / curl (Keystone/Nova/…) |
| 09 Edge dataplane | FluxVM Network Fabric via Fabric API / CLI / UI |
| 15 AI Workloads | OpenAI inference: Janus or NVIDIA, gateway, MIG (Beta) |
| 16 Keep workstation | Keep mode, signed policy, cockpit, browser |
| 17 Keep PDF brief | One-click PDF → brief.md, 0 CONNECT |
| OpenStack Compatibility reference | Endpoints, public URL, limitations |
| Proven infrastructure | Compatibility matrix, SLOs, chaos, upgrade/rollback (#14–#17) |
Guides
CLI Guides
| Document | Description |
|---|---|
| fabricctl Reference | Full CLI command reference with examples |
| zyvor-fabricd-ctl Operations | Deployment, management, and maintenance commands |
| Declarative Configuration | Define VM infrastructure as YAML with fabricctl apply |
| Shell Completions | Enable tab completion for bash |
Operations Guides
| Document | Description |
|---|---|
| Production Deployment | Hardening, TLS, monitoring, and logging for production |
| Capacity Planning | Resource estimation and scaling guidelines |
| Backup Strategy | Backup types, retention policies, and disaster recovery |
| Monitoring Setup | Prometheus integration, alerting, and dashboards |
| Troubleshooting | Common issues, log analysis, and diagnostic steps |
| Upgrade Procedures | Rolling upgrades and version migration |
| Security Hardening | Firewall rules, RBAC policies, and credential rotation |
Decision Support
| Document | Description |
|---|---|
| Zyvor Fabric vs. Proxmox | Feature and architecture comparison |
| Zyvor Fabric vs. OpenStack | Scope, complexity, and deployment comparison |
| Zyvor Fabric vs. libvirt | Integration model and management layer comparison |
| Storage Backend Selection | Choosing between Local, NFS, LVM, ZFS, and Ceph |
| Networking Architecture | Overlay vs. flat networking decisions |
Features
Detailed documentation for each major feature area.
VM Lifecycle
| Document | Description |
|---|---|
| VM Creation | CreateVMRequest fields, validation rules, defaults |
| VM Start Options | VMStartOptions reference -- KVM, TPM, Secure Boot, networking |
| VM States | State machine: Stopped, Starting, Running, Paused, Stopping, Failed |
| VM Cloning | Full and linked clones with CoW support |
| VM Templates | Create, manage, and deploy from templates |
| VM Profiles | Instance types and resource presets |
| VM Import | Import from VMDK, VDI, VHD formats |
| OVA/OVF Export | Export VMs to OVA/OVF format for portability |
| Declarative Specs | YAML-based VM definitions with fabricctl apply |
| VM Checkpoints | Create and restore in-memory checkpoints |
| VM Forking | Fork a running VM for testing |
Container Workloads
| Document | Description |
|---|---|
| Container Groups | Kubernetes-style Pod groups backed by FluxVM Secure Containers -- tenant scoping, image pull secrets, health probes, NetworkPolicy, quotas/billing, backup/restore |
Storage
| Document | Description |
|---|---|
| Storage Overview | Storage architecture and backend selection |
| Local Storage | File-based qcow2/raw storage pools |
| NFS Storage | NFS Storage Guide |
| LVM Storage | Logical Volume Manager pools and thin provisioning |
| ZFS Storage | ZFS pools, datasets, and replication |
| Ceph/RBD Storage | Ceph cluster integration and RBD image management |
| Volume Management | CRUD, attach/detach, resize, and clone volumes |
| Snapshots | Create, revert, delete, and tree-view snapshots |
| Distributed Storage | Cross-node storage pools, migrations, and policies |
| Datastore Clusters | Storage DRS and placement recommendations |
| iSCSI Storage | iSCSI target discovery, login, and LUN management |
| Cloud Images | Built-in catalog download (Ubuntu, Fedora, Debian, Alma) |
| ISO Management | Download, list, and manage ISO images |
| Online Disk Resize | Grow VM disks without downtime |
Networking
| Document | Description |
|---|---|
| Networking Overview | Network architecture, SDN stack, and VM edge dataplane |
| VM edge dataplane (Network Fabric schema v4) | Fabric proxy of FluxVM TC/eBPF — API, Web, CLI, lab UX |
| Service Fabric v6 (Maglev VIP LB) | FluxVM Maglev/NAT/DSR/EDT/flows/policy + Fabric leases, HA, FRR/BIRD |
| AI Workloads (Beta) | GPU VMs, Janus lab upstream, PCI MIG flag, admit webhook, fabric-inference JWT, revisioned rollouts, gateway limits, Maglev, Terraform, operator CRDs |
| User: VM Dataplane | Console walkthrough for Status / Policy / Effective / Stats / Flows |
| User: Edge Dataplane | Cluster groups / CNP / Maglev Services / health / observe console |
| Tutorial 09: Edge Dataplane | End-to-end lab (API + CLI + UX) |
| Edge dataplane series | Short labs: groups, CNP, FQDN, UX |
| Networking (netlink) | Bridges, VLANs, bonds, taps, macvtaps, VXLANs, SR-IOV |
| Network Policies | Cilium-style label-based ingress/egress rules |
| VM Firewall | Per-VM firewall profiles and zones via nftables |
| Service Mesh | Virtual IP load balancing (round-robin, least-conn, IP-hash) |
| Traffic Shaping | QoS: guaranteed rate, max rate, burst, priority-based bandwidth |
| DNS Policies | Zone management, upstream servers, domain blocking |
| VPN Mesh | WireGuard tunnels: point-to-point, hub-spoke, full-mesh |
| NAT Gateway | Masquerade, SNAT, DNAT, hairpin NAT |
| Packet Mirror | Traffic capture and debugging |
| Network Monitor | Per-VM bandwidth tracking with threshold alerts |
| Floating IPs | Virtual IP allocation and VM assignment |
| DHCP Servers | Per-bridge dnsmasq-managed DHCP servers |
| DHCP Server | Built-in DHCP server on bridge interfaces |
| Port Forwarding | NAT-based port forwarding rules |
Security and Identity
| Document | Description |
|---|---|
| Security Overview | Security architecture and audit history |
| Authentication | PAM + JWT authentication flow |
| RBAC | Admin, User, Viewer roles and endpoint permissions |
| API Keys | Service-to-service authentication tokens |
| OIDC / External Auth | LDAP and OIDC/OAuth2 SSO (PKCE, JWKS-verified id_token) |
| SCIM Provisioning | SCIM 2.0 lifecycle provisioning and group-to-role sync for Entra ID / Okta |
| OpenStack Compatibility | Experimental Keystone/Nova/Glance/Neutron/Cinder façade on the same daemon port |
| Audit Logging | Structured audit logs with JSON/CSV export |
| Encryption | VM disk encryption with key management providers |
| TLS/HTTPS | Certificate management and self-signed TLS generation |
| Multi-Tenancy | Project isolation with member roles and quotas |
| 2FA/MFA Authentication | TOTP-based two-factor authentication for user accounts |
| Secrets Management | Secure storage and retrieval of credentials and sensitive data |
| Compliance Scanning | Automated compliance profile scanning and reporting |
| JWT Revocation | Per-token revocation via JTI blocklist |
| Credential Management | systemd credentials, SSH keys, and cloud-init secrets |
High Availability
| Document | Description |
|---|---|
| DRS | Distributed Resource Scheduling and placement |
| Affinity Rules | VM-to-VM and VM-to-host affinity/anti-affinity |
| Host Maintenance Evacuation | Preflight-checked workload evacuation before a host enters maintenance |
| Fabric Doctor | Production host preflight checks and safe support bundles (tools/fabric-doctor) |
| Fault Tolerance | Automatic failover, fencing, and recovery |
| Live Migration | Disk-copy: iterative rsync pre-copy + cutover; native FluxVM transport = preview (receivers) |
| Site Recovery | Failover/reprotect workflows for disaster recovery |
| Resource Overcommit | CPU/memory/storage overcommit policies |
| Split-Brain Protection | Quorum-based fencing to prevent split-brain in clusters |
| Auto-Scaling | Metric-based scaling policies and events |
Monitoring and Automation
| Document | Description |
|---|---|
| Prometheus Metrics | /metrics endpoint and metric catalog |
| Analytics Dashboard | Historical performance data and insights |
| Notifications | Email, Slack, Webhook, Microsoft Teams channels |
| Schedules | Once, daily, weekly VM operations scheduling |
| Backup Automation | Retention policies, incremental backups, systemd timers |
| Resource Quotas | Per-user and per-project resource limits |
| Optimization | Resource optimization recommendations |
| Log Aggregation | Centralized VM log collection, search, and streaming |
| Health Checks | Deep health check: API, disk, DB, credentials, KVM |
Console Access
| Document | Description |
|---|---|
| WebSocket Console | Browser-based terminal via xterm.js |
| VNC Console | Graphical console via noVNC proxy |
| SPICE Display | SPICE protocol support for high-performance remote display |
| Console Modes | Interactive, read-only, native, GUI |
Advanced Virtualization
| Document | Description |
|---|---|
| GPU Passthrough | Generic PCI/VFIO passthrough for NVIDIA and AMD GPUs (no vGPU/Intel GVT-g) |
| CPU/NUMA Optimization | CPU pinning, NUMA topology, hugepages |
| KSM Memory Dedup | Kernel Same-page Merging configuration |
| Nested Virtualization | Running VMs inside VMs |
| Hotplug | CPU, memory, disk, and NIC hot-add/remove |
| Firmware | UEFI, Secure Boot, NVRAM management |
| TPM | vTPM support via swtpm |
| Direct Kernel Boot | Boot from kernel + initrd without bootloader |
| Bind Mounts | Host-to-VM filesystem sharing |
| USB Passthrough | Host USB device passthrough to VMs |
| User Namespaces | Private user mapping for isolation |
Architecture
| Document | Description |
|---|---|
| Architecture Overview | System architecture and component diagram |
| Crate Structure | 53 backend crates and their responsibilities |
| Data Model | VM, VMStartOptions, VMMetrics, VMPressure |
| Driver Model | VMDriver and ResourceStatsDriver traits |
| State Store | SQLite-based persistent state management |
| Event System | Broadcast channels and SSE event streaming |
| Background Tasks | Reconciliation loops and schedulers |
| Plugin System | Plugin registry and extension points |
Reference
API Reference
| Document | Description |
|---|---|
| API Overview | Authentication, pagination, error format |
| VM Endpoints | CRUD, lifecycle, metrics, cloud-init |
| Storage Endpoints | Pool and volume management |
| Network Endpoints | Bridges/VLANs/bonds via netlink, policies, firewall, mesh |
| Security Endpoints | Auth, audit, encryption, tenants |
| System Endpoints | CPU topology, NUMA, memory, firmware |
| Enterprise Endpoints | Datacenters, clusters, hosts, DRS |
| Machine Endpoints | VM driver: FluxVM (no systemd dependency) |
| Monitoring Endpoints | Analytics, events, notifications, schedules |
| Backup Endpoints | Backup CRUD, policies, restore |
| Billing Endpoints | Usage tracking, pricing, and invoicing |
| WebSocket Endpoints | Console, VNC, event streaming |
CLI Reference
| Command | Description |
|---|---|
fabricctl / fabricctl --help | Cilium-style grouped help with emoji section markers (Basic, Dataplane, Networking, Meta, …) |
fabricctl list | List VMs (paginated {items, total, …} from GET /api/vms); set FABRIC_URL + FABRIC_TOKEN or ZYVOR_FABRIC_* for HTTPS labs |
fabricctl create | Create a new VM |
fabricctl start | Start a stopped VM |
fabricctl stop | Stop a running VM |
fabricctl restart | Restart a VM |
fabricctl delete | Delete a VM |
fabricctl apply | Apply declarative YAML specification |
fabricctl policy | Manage network policies |
fabricctl dataplane | VM edge dataplane (policy, flows, hubble, CNP, …) |
fabricctl status | Fabric API + dataplane health checklist |
fabricctl config | Show effective server/token/color settings |
fabricctl completion | Generate bash/zsh/fish completion scripts |
fabricctl ceph | Ceph storage management |
fabricctl metrics | Get VM metrics |
Global flags: -o table|json|yaml, --color auto|always|never, --server, --token.
VM console/VNC access is Web/REST-only (GET /ws/console/:name, /ws/vnc/:name) — there is no fabricctl console command.
zyvor-fabricd-ctl Reference
| Command | Description |
|---|---|
zyvor-fabricd-ctl deploy | Full deployment (deps + build + install + start) |
zyvor-fabricd-ctl deps | Install system dependencies |
zyvor-fabricd-ctl build | Build from source |
zyvor-fabricd-ctl install | Install binaries and systemd units |
zyvor-fabricd-ctl start | Start the zyvor-fabricd service |
zyvor-fabricd-ctl stop | Stop the zyvor-fabricd service |
zyvor-fabricd-ctl restart | Restart the service |
zyvor-fabricd-ctl status | Show service status |
zyvor-fabricd-ctl logs | Follow service logs |
zyvor-fabricd-ctl verify | Post-install smoke test |
zyvor-fabricd-ctl health | Deep health check |
zyvor-fabricd-ctl password | Read the admin password |
zyvor-fabricd-ctl doctor | System readiness check |
zyvor-fabricd-ctl tls | Generate self-signed TLS certificate |
zyvor-fabricd-ctl upgrade | Git pull + reinstall |
zyvor-fabricd-ctl uninstall | Remove everything |
zyvor-fabricd-ctl billing | View billing and usage reports |
zyvor-fabricd-ctl backup now | Trigger immediate backup |
zyvor-fabricd-ctl backup enable | Enable daily backup timer |
zyvor-fabricd-ctl backup status | Show backup timer and storage info |
Rust SDK
| Item | Description |
|---|---|
zyvor-fabric-sdk | Typed Rust SDK for the Zyvor Fabric API with async client |
| Authentication | Login, token refresh, and 2FA helpers |
| VM Operations | Create, start, stop, delete, clone, and snapshot VMs |
| Storage / Networking | Pool, volume, bridge, VLAN, and firewall management |
| Streaming | WebSocket console attach and SSE event subscriptions |
Deployment
| Document | Description |
|---|---|
| Kubernetes | Run fabricd + FluxVM as privileged DaemonSets; Helm; lab ./scripts/deploy k8s |
| Docker / Podman | Local eval with compose (hostNetwork + KVM) |
| Single Server | One-node deployment for development and small teams |
| Multi-Node Cluster | HA deployment with shared storage and etcd |
| Kubernetes Operator | VMs as CRDs with the zyvor-fabricd operator (operator/) |
| Terraform Provider | Declarative provisioning with plan/apply |
| Edge Deployment | Lightweight single-node deployment for edge locations |
| Air-Gapped Install | Offline installation without internet access |
Development
| Document | Description |
|---|---|
| Development Setup | Rust toolchain, IDE, and local development |
| Build and Test | cargo check, cargo test, CI pipeline |
| Crate Map | 53 crates and their dependencies |
| Adding an API Endpoint | Step-by-step guide for new endpoints |
| Adding a Storage Backend | Driver trait implementation guide |
| Adding a Network Feature | Integration with the netlink-based networking crate |
| Code Style | Formatting, naming, error handling conventions |
| Security Guidelines | Input validation, path traversal prevention, audit |
Quick Reference
See also the standalone quick-reference docs: FAQ · Glossary · Quick reference cheatsheet.
API Endpoint Categories
The REST API is organized into the following endpoint groups:
| Category | Prefix | Endpoints | Description |
|---|---|---|---|
| Authentication | /api/auth/ | 6 | Login, 2FA/TOTP setup, and session management |
| SCIM Identity | /api/identity/scim/, /scim/v2/ | 21 | Provisioning profiles/tokens (JWT) plus SCIM Users/Groups (bearer token) |
| OpenStack Compatibility | /identity, /compute, /image, /network, /volume | — | Experimental OpenStack wire protocol (catalog from public_url) |
| VM Lifecycle | /api/vms/ | 12 | CRUD, start, stop, restart, pause, resume, clone |
| VM Advanced | /api/vms/{name}/ | 20+ | Hotplug, checkpoints, fork, disk resize, firmware |
| Snapshots | /api/vms/{name}/snapshots/ | 5 | Create, list, get, delete, revert |
| Cloud-Init | /api/vms/{name}/cloud-init | 1 | Configure cloud-init for a VM |
| Storage Pools | /api/storage/pools/ | 14 | Pool CRUD, health, stats, refresh |
| Volumes | /api/storage/pools/{name}/volumes/ | 6 | Volume CRUD, resize, attach, detach |
| Distributed Storage | /api/distributed-storage/ | 18 | Cross-node pools, migrations, policies |
| Networking | /api/networkd/ | 35+ | Bridges, VLANs, bonds, taps, VXLANs, SR-IOV (netlink-based) |
| Network Policies | /api/network-policies/ | 5 | Cilium-style ingress/egress rules |
| VM Firewall | /api/vm-firewall/ | 8+ | Per-VM firewall profiles and zones |
| Service Mesh | /api/service-mesh/ | 10+ | Virtual IP and load balancing |
| Traffic Shaping | /api/traffic-shaping/ | 8+ | QoS and bandwidth management |
| DNS Policies | /api/dns-policies/ | 6+ | Zone management and blocking |
| VPN Mesh | /api/vpn-mesh/ | 6+ | WireGuard tunnel management |
| NAT Gateway | /api/nat-gateway/ | 6+ | Masquerade, SNAT, DNAT |
| Packet Mirror | /api/packet-mirror/ | 4+ | Traffic capture |
| Net Monitor | /api/net-monitor/ | 4+ | Bandwidth tracking and alerts |
| Floating IPs | /api/floating-ips/ | 4 | Virtual IP allocation |
| DHCP | /api/dhcp-servers/ | 2 | Per-bridge dnsmasq-managed DHCP servers |
| DNS | /api/dns/ | 4 | DNS configuration |
| Encryption | /api/encryption/ | 11 | Key providers, policies, VM encryption |
| Backups | /api/backups/ | 11 | Backup CRUD, policies, restore, stats |
| Schedules | /api/schedules/ | 9 | Timed VM operations |
| Quotas | /api/quotas/ | 8 | Resource quotas and usage |
| Notifications | /api/notifications/ | 11 | Multi-channel alerting |
| Audit | /api/audit/ | 4 | Audit logs and export |
| Analytics | /api/analytics/ | 6 | Performance data and insights |
| Templates | /api/templates/ | 5 | VM templates and deployment |
| Profiles | /api/profiles/ | 3 | Instance type presets |
| Images | /api/images/ | 10 | Image build, cloud download, ISO, import |
| Migrations | /api/migrations/ | 3 | Live VM migration |
| Events | /api/events/ | 2 | Event list and SSE stream |
| System | /api/system/ | 12 | CPU, NUMA, memory, hugepages, optimization |
| Firmware | /api/vms/{name}/firmware/ | 5 | UEFI, Secure Boot, NVRAM |
| Datacenters | /api/datacenters/ | 5 | Datacenter management |
| Clusters | /api/clusters/ | 5 | Cluster management and health |
| Hosts | /api/hosts/ | 8 | Host registration and maintenance |
| DRS | /api/drs/ | 9 | Scheduling, placement, affinity rules |
| Resource Pools | /api/resource-pools/ | 7 | Resource pool management |
| Zones | /api/zones/ | 3 | Availability zones |
| Spot Instances | /api/spot-instances/ | 3 | Spot VM management |
| Tenants | /api/tenants/ | varies | Multi-tenancy and projects |
| Settings | /api/settings | 2 | Global settings |
| Plugins | /api/plugins | 1 | Plugin registry |
| OVA Export | /api/vms/{name}/export/ | 2 | OVA/OVF export |
| Secrets | /api/secrets/ | 5 | Secrets and credential management |
| Compliance | /api/compliance/ | 6 | Compliance profile scanning and results |
| Billing | /api/billing/ | 6 | Usage tracking, pricing, and invoicing |
| Logs | /api/logs/ | 4 | Centralized log aggregation and search |
| iSCSI | /api/iscsi/ | 5 | iSCSI target discovery and session management |
| USB | /api/usb/ | 4 | USB device listing and passthrough |
| SPICE | /api/vms/{name}/spice/ | 2 | SPICE display connection |
| Declarative | /api/vms/apply | 2 | YAML spec apply and export |
| Auto-Scale | /api/autoscale/ | 4 | Scaling policies and events |
| WebSocket | /api/ws/ | 3 | Console, VNC, events |
Default Ports and Paths
| Item | Default |
|---|---|
| API listen address | 127.0.0.1:9095 |
| Config file | /etc/zyvor-fabricd/zyvor-fabricd.toml |
| State directory | /var/lib/zyvor-fabricd/ |
| Image directory | /var/lib/zyvor-fabricd/images/ |
| Auth database | /var/lib/zyvor-fabricd/auth.db |
| Cloud-init directory | /var/lib/zyvor-fabricd/cloud-init/ |
| Storage pools | /var/lib/zyvor-fabricd/storage/ |
| JWT secret file | /var/lib/zyvor-fabricd/.jwt_secret |
| Admin password file | /var/lib/zyvor-fabricd/.admin_password |
| networkd config dir | /etc/systemd/network/ |
| networkd file prefix | 50-Zyvor Fabric- |
| Network bridge | br0 |
VM Resource Limits
| Resource | Minimum | Maximum | Default |
|---|---|---|---|
| CPUs | 1 | 256 | -- |
| Memory | 128 MB | 1,048,576 MB (1 TB) | -- |
| Disk | 1 GB | 65,536 GB (64 TB) | 20 GB |
VM States
| State | Description |
|---|---|
stopped | VM is not running |
starting | VM is in the process of booting |
running | VM is running and accessible |
paused | VM is suspended in memory |
stopping | VM is in the process of shutting down |
failed | VM failed to start or encountered an error |
unknown | VM state could not be determined |
RBAC Roles
| Role | Read | Write | Admin |
|---|---|---|---|
| Viewer | Yes | No | No |
| User | Yes | Yes | No |
| Admin | Yes | Yes | Yes |
Environment Variables
| Variable | Description |
|---|---|
ZYVOR_FABRICD_LISTEN | Override daemon.listen (e.g. 0.0.0.0:9095) |
ZYVOR_FABRICD_PUBLIC_URL | External base URL for OpenStack catalog / clients |
ZYVOR_FABRICD_CONFIG | Override config file path |
ZYVOR_FABRICD_LOG_LEVEL | Override log level |
ZYVOR_FABRICD_JWT_SECRET | Override JWT signing secret |
ZYVOR_FABRICD_ADMIN_PASSWORD | Override default admin password |
ZYVOR_FABRICD_BACKUP_DIR | Override backup directory |
ZYVOR_FABRICD_BACKUP_RETAIN | Override backup retention count |
ZYVOR_FABRICD_BACKUP_TYPE | Override backup type |
AI Workloads (Beta)
OpenAI-compatible inference on dedicated NVIDIA GPU VMs: ModelArtifact /
InferenceProfile / InferenceDeployment / InferenceEndpoint, revisioned
rollouts, Maglev weighting, gateway rate limits, an API-key (or
fabric-inference JWT) OpenAI gateway at /api/ai/openai/{endpoint}, and an
optional Janus stand-in when the host has no NVIDIA GPU. Console: /app/ai.
Single-cluster Beta (not GA). Multi-site HA store stays Preview.
| Document | Description |
|---|---|
| ai-workloads.md | Full reference: REST, CLI, Janus, PCI MIG, admit, Raft, runtimes |
| Tutorial 15 | Operator how-to: model → chat → MIG |
| scripts/bake-ai-vllm-image.sh | Bake CUDA + vLLM qcow2 |
| scripts/smoke-ai-janus-lab.sh | Lab smoke: health, chat, MIG, admit |
| terraform-provider/examples/ai-workloads | Terraform model/profile/deployment/endpoint |
| operator/examples/ai-inference-deployment.yaml | ModelArtifact + InferenceDeployment CRs |
| operator chart admissionWebhook | Opt-in ValidatingWebhook for InferenceDeployment |
| Agent Runtime ↔ Fabric | kind: fabric credentials for on-prem models |
Agent Runtime
Standalone component: durable agent sessions in a FluxVM sandbox, or on a GitHub runner through a FluxVM stand-in. A session runs a JavaScript worker or a claude, codex, or gemini CLI. Provider keys stay on the host egress broker. The crate also accepts cron schedules, signed webhooks, bounded loops, operator approvals, agent-to-agent delegation, and a small MCP server. Lives in agent-runtime/ (Rust daemon), sdk/agent-runtime/ (SDK + fabric-agent CLI), and examples/agent-runtime/ — outside the backend/ Cargo workspace.
| Document | Description |
|---|---|
| agent-runtime/README.md | Deploy/run, harness, schedules, MCP, HTTP API, warm pools, session CI |
| Tutorial 11 | Build a Node.js FluxVM template and drive a session |
| Tutorial 13 | Build a bundle and deploy it from the web console |
| Tutorial 14 | Use cases: ops, coding, research, cron, webhook, loop, approval, handoff, IDE chat, fan-out |
| examples/agent-runtime/README.md | Ops health check, hello-go, warm pool, and fan-out |
Integrations
- integrations/README.md — Machina, Terraform, operator, Ansible, OpenStack-compat
- openstack-compat.md — drive Fabric with
openstack/ Terraform OpenStack / Ansibleos_* - DEVOPS.md — CI/CD, GitOps, Terraform, probes, promotion with FluxVM
- examples/devops/ — copy-paste pipelines and gates
- NAMING.md — Clone URL and product naming
Product Positioning
- docs/POSITIONING.md — Zyvor Fabric vs. Machina, naming model, messaging
- client-presentations/ — Client-facing HTML slide decks
Client Presentations
| Document | Description |
|---|---|
| Product Overview | Comprehensive product overview with feature matrix |
| Product Overview (PDF) | Printable product overview |
| Security Audit Report | Full security audit report |
| Security Audit Report (PDF) | Printable security audit report |
This index is maintained alongside the codebase. For the latest information, refer to the source code and inline documentation.