> For the complete documentation index, see [llms.txt](https://docs.platform9.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.platform9.com/private-cloud-director/introduction/technical-data-sheet.md).
# Technical Data Sheet
Specification-level summary of the platform: architecture, host stack, guest virtual machines, availability, networking, storage, security, and operations.
space.vars.product\_name (space.vars.product\_acronym) is an enterprise private cloud platform that lets IT and DevOps administrators manage virtualized and Kubernetes infrastructure at scale, on their existing storage and server hardware. It is built on the open source KVM hypervisor and combines open source technologies for virtualization and container management with Platform9's own features, so VMs and containers are managed side by side.
* **Virtualization-first management:** a familiar interface for virtualization administrators, with the traditional virtualization features they rely on.
* **Enterprise virtualization features built in:** VM High Availability, Dynamic Resource Rebalancing, VM cloning and snapshots, and VM affinity and anti-affinity rules.
* **Existing storage and servers:** integrates with a wide range of enterprise storage arrays and with existing server hardware, without requiring a refresh.
* **Kubernetes on the same platform:** Kubernetes clusters run on the same management plane, hosts and tenants as VMs, and are operated from one UI and one API.
* **Two deployment models:** a SaaS management plane operated by Platform9, or a Self-Hosted management plane in the customer's data center. Workloads run on the customer's hardware in both.
* **Migration from VMware:** vJailbreak, a free Platform9 tool, converts vSphere clusters to space.vars.product\_acronym clusters in place, in a rolling manner.
This data sheet specifies the platform: architecture, host stack, guest virtual machines, availability, networking, storage, security and operations. Every fact comes from this documentation, and each section links its source pages.
## Platform Overview and Architecture
space.vars.product\_acronym separates a management plane, which holds the core services, the API and the UI, from a data plane of customer-supplied physical hosts that run the KVM hypervisor and the VMs. Workloads always run in the customer's data center on the customer's hardware, wherever the management plane runs. Applications run unmodified inside their guest operating system.
| Layer | Component | Function |
| --------------------- | ------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Hypervisor | KVM, with QEMU and libvirt | Runs each workload as a VM; KVM runs in the Linux kernel on the host hardware |
| Host operating system | Platform9 OS (Rocky Linux from CIQ 10.2) or Ubuntu 24.04 LTS | Base OS on each hypervisor host (see Hypervisor and Host Stack) |
| Host agent | Platform9 host agent | Installs the software each host role needs, reports host health and carries out upgrades; it is the only path between the management plane and the hosts |
| Compute | Compute Service | Provisions VMs, decides which host each one runs on and manages the VM lifecycle |
| Cluster automation | VM High Availability (VM HA) and Dynamic Resource Rebalancing (DRR) | Recovers VMs after a host failure; live migrates VMs to relieve CPU and memory imbalance |
| Networking | Networking Service, on OVN and Open vSwitch | Virtual networks, subnets, routers, security groups and load balancing |
| Block storage | Persistent Storage Service | Persistent volumes for VMs; integrates the customer's storage backends |
| Images | Image Library Service | Manages VM images and snapshots and serves them to hypervisors during provisioning |
| Identity | Identity Service | Authenticates users and API clients and authorizes each request against domains, tenants and roles |
| Platform services | Load Balancer as a Service and DNS as a Service | Infrastructure services that tenants provision for their applications |
| Kubernetes | Kubernetes clusters with hosted control planes | Worker nodes run as VMs on a virtualized cluster, or on bare metal nodes (beta). Control planes run as pods in the management plane, and cluster lifecycle is declarative through Cluster API |
| Management plane | SaaS or Self-Hosted | Holds cloud state and makes placement decisions; it does not run workloads |
space.vars.product\_acronym ships in two commercial deployment models that differ only in where the management plane runs and who operates it. In the SaaS model Platform9 hosts and operates it, and each host connects outbound on port 443. In the Self-Hosted model the customer runs it on its own hardware, including in air-gapped environments. The data plane is identical in both.
Sources: [Overview](/private-cloud-director/introduction/readme.md), [Design Principles](/private-cloud-director/introduction/design-principles.md), [Architecture Overview](/private-cloud-director/introduction/architecture-overview.md), [Kubernetes Overview](/private-cloud-director/kubernetes-clusters/k8s-overview.md), [Pre-requisites](/private-cloud-director/getting-started/pre-requisites.md), [Load Balancer as a Service](/private-cloud-director/virtualized-networking/load-balancer.md).
## Hypervisor and Host Stack
space.vars.product\_acronym runs on x86 servers with hardware virtualization enabled and works with existing server hardware without requiring a refresh.
| Item | Specification |
| ---------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Hypervisor | KVM, with QEMU and libvirt |
| Supported host OS | Platform9 OS (Rocky Linux from CIQ 10.2), installed from the Platform9 OS ISO; Ubuntu 24.04 LTS |
| Deprecated host OS | Ubuntu 22.04 LTS: deprecated since the August 2026 release, end of support with the April 2027 release. See the [Host Operating System Support Matrix](/private-cloud-director/getting-started/host-operating-system-support-matrix.md) for the full timeline |
| Host OS hardening | Platform9 OS ships hardened to the CIS RHEL 10 Benchmark v1.0.1, Level 1 Server profile, and pulls packages only from the Platform9 validated repository. Ubuntu hosts are installed and maintained by the customer |
| Server architecture | x86 only |
| CPU | Intel with VT-x and Intel 64, or AMD with AMD-V and AMD64, enabled in firmware; must meet the CPU model prerequisites |
| Minimum per host | 8 vCPUs, 16 GB RAM, 250 GB local storage (100 GB when VM root disks use block storage volumes or shared storage) |
| Hardware compatibility | Platform9 does not publish a server or I/O device compatibility list. Any x86 server that meets the [Pre-requisites](/private-cloud-director/getting-started/pre-requisites.md) is supported. Block storage arrays must use a certified driver (see Storage and Data Protection) |
| Host networking | At least one network interface; four recommended as two bonded pairs (LACP or active-backup). Management, image library, VM console and virtual network tunnel traffic can use separate interfaces |
| Time synchronization | An NTP service on every host (chronyd, ntpd, ntp or systemd-timesyncd) |
| Mixed host OS clusters | All hosts in a cluster should run the same OS version. Live migration and VM evacuation are not supported between Ubuntu 22.04 and 24.04 hosts, or between Ubuntu and Platform9 OS hosts; cold migration from Ubuntu to Platform9 OS hosts is supported |
Sources: [Pre-requisites](/private-cloud-director/getting-started/pre-requisites.md), [Host Operating System Support Matrix](/private-cloud-director/getting-started/host-operating-system-support-matrix.md), [Platform9 OS CIS Hardening](/private-cloud-director/getting-started/platform9-os/platform9-os-cis-hardening.md), [Networking Overview](/private-cloud-director/virtualized-networking/networking-overview.md), [CPU Model](/private-cloud-director/virtualized-clusters/add-hosts-virtualized-cluster/cpu-model.md).
## Guest Virtual Machine Specification
A guest on space.vars.product\_acronym sees an x86-64 virtual machine whose chipset, firmware, disk bus and network adapter model are set through image properties. Paravirtualized VirtIO disk and network devices are the recommended configuration.
| Item | Specification |
| --------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Guest operating systems | Windows and Linux guests. The versions qualified with the QEMU guest agent are listed in [QEMU Guest Agent](/private-cloud-director/virtualized-clusters/virtualmachine/qemu-guest-agent.md) |
| CPU | Every host in a virtualized cluster exposes the same CPU model to its VMs, which keeps the instruction set consistent for live migration |
| Machine type | `q35`, or `pc` (i440fx, legacy), set with the `hw_machine_type` image property |
| Firmware | BIOS or UEFI, set with the `hw_firmware_type` image property |
| Secure Boot | Enabled with the `os_secure_boot` image property |
| Virtual TPM | vTPM 2.0 with the `tpm-tis` or `tpm-crb` model, enabled through image metadata. Live migration and VM HA for vTPM VMs need the vTPM state directory on shared storage |
| Disk | VirtIO disk bus; up to 8 virtual disks per VM (root plus attached volumes); boot volume up to 4 TB |
| Network | VirtIO network adapter; multiple ports per VM |
| Guest agent | QEMU guest agent, installed by the customer. On Windows it installs from the virtio-win guest tools together with the VSS provider used for application-consistent snapshots. space.vars.product\_acronym does not use the agent for VM health monitoring |
| Changes without power-off | Hot-add CPU and memory (increase only; the VM must be created with a zero-size flavor); extend an in-use volume, then extend the file system in the guest; attach or detach a data volume |
| Changes that need power-off | Flavor resize, which powers the VM off and recreates it; reducing CPU or memory; the Mount ISO action |
Sources: [QEMU Guest Agent](/private-cloud-director/virtualized-clusters/virtualmachine/qemu-guest-agent.md), [Performance Tuning](/private-cloud-director/virtualized-clusters/performance-tuning.md), [Image Properties](/private-cloud-director/images-and-image-library/image-properties.md), [Virtual TPM](/private-cloud-director/virtualized-clusters/virtual-tpm.md), [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md), [VM Hot Add CPU or Memory](/private-cloud-director/virtualized-clusters/virtualmachine/vm-hot-add-cpu-or-memory.md), [Virtual Machine Actions](/private-cloud-director/virtualized-clusters/virtualmachine/virtual-machine-actions.md), [Resize a Volume](/private-cloud-director/tutorials/resize-a-volume.md), [Mount an ISO on a Running VM](/private-cloud-director/virtualized-clusters/virtualmachine/mount-iso-on-a-running-vm.md).
## Availability and Resiliency
space.vars.product\_acronym restarts VMs on healthy hosts after a host failure and live migrates running VMs between hosts for maintenance and load balancing. Both behaviors are configured per virtualized cluster.
| Capability | Behavior | VMware equivalent |
| ---------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------ |
| VM High Availability | Monitors host health in an HA-enabled cluster and restarts a failed host's VMs on healthy hosts in the same cluster. Needs at least two hosts and shared storage (volume-backed root disks or shared ephemeral storage). Response time from host failure to the first VM evacuation is about 3.5 minutes by default, configurable to about 2 minutes or 1 minute | VMware High Availability |
| Live migration | Moves a running VM to another host with a brief pause, typically milliseconds to seconds. Volume-backed VMs move without a storage copy. Source and destination hosts must run the same OS version | vMotion |
| Cold migration | Moves a powered-off or suspended VM between hosts | Cold migration |
| Storage live migration | Migrates a VM's storage as part of live migration; volume retype moves a volume across storage backends | Storage vMotion |
| Dynamic Resource Rebalancing | Evaluates the cluster every 20 minutes by default (10 or 30 minutes configurable) and live migrates VMs to relieve overloaded hosts | DRS |
| Placement controls | Host aggregates direct VMs to hosts that share a property. Hard and soft affinity and anti-affinity rules keep VMs together or apart, and VM HA applies them when it restarts a VM | Host groups and affinity rules |
| Stretched clusters | Splits one cluster across two sites as two failure domains, with a soft affinity per VM; VM HA recovers workloads on the surviving site if a whole site fails | vSphere Metro Storage Cluster |
| Host maintenance | Maintenance mode live migrates all running VMs off a host and marks it unschedulable | Maintenance mode |
| Platform upgrades | Running VMs are not affected by a management plane upgrade; the UI and API are unavailable while a Self-Hosted management plane upgrades. Hosts are drained with maintenance mode before they upgrade. A VM left on a host during its upgrade loses network packets for about 4 to 6 seconds | vCenter and ESXi upgrades |
Tested limits: 100 hypervisors per cluster, 2 failure domains per cluster and 5 concurrent VM migrations per region.
Sources: [VM High Availability](/private-cloud-director/virtualized-clusters/virtualized-cluster/virtual-machine-high-availability-vm-ha.md), [Virtual Machine Migration](/private-cloud-director/virtualized-clusters/virtualmachine/vm-migration.md), [Storage Live Migration](/private-cloud-director/storage/storage-live-migration.md), [Dynamic Resource Rebalancing](/private-cloud-director/virtualized-clusters/virtualized-cluster/dynamic-resource-rebalancing-drr.md), [VM Affinity Anti-Affinity Rules](/private-cloud-director/virtualized-clusters/virtualmachine/vm-affinity-anti-affinity-rules.md), [Stretched Clusters](/private-cloud-director/virtualized-clusters/stretched-clusters.md), [Maintenance Mode](/private-cloud-director/virtualized-clusters/add-hosts-virtualized-cluster/maintenance-mode.md), [Host Upgrade Runbook](/private-cloud-director/upgrade/host-upgrade-runbook.md), [Upgrade Management Plane and Hosts](/private-cloud-director/getting-started/self-hosted/upgrade.md), [Transition from VMware](/private-cloud-director/getting-started/transition-from-vmware.md), [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md).
## Networking
VMs attach directly to existing data center VLANs through physical networks, where they share a Layer 2 segment with the servers, firewalls and other network infrastructure already on that VLAN. Software-defined virtual networks are available where isolation or self-service is needed.
| Capability | Support |
| ---------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Attach VMs to existing VLANs | Physical networks (called provider networks in the API), flat or VLAN (802.1Q), map to existing Layer 2 networks in the data center; DHCP and metadata services are optional |
| Virtual networks | VLAN, VXLAN or Geneve networks with routers; Distributed Virtual Routing is the default mode |
| Layer 2 networks | Connect VMs without managed IP, DHCP or routing services; up to 500 ports per Layer 2 network |
| Multiple network interfaces per VM | Multiple ports per VM, including several ports from the same physical network |
| Addressing | IPv4 and dual-stack (IPv4 plus IPv6) ports; a port can be created in advance with a specific IP address |
| Security groups | Per-VM inbound and outbound rules; traffic that matches no rule is denied. Up to 10 security groups per port and 100 rules per group |
| Load balancing | Built-in Load Balancer as a Service, implemented in the OVN distributed router with no load balancer VMs |
| DNS | DNS as a Service and external DNS integration |
| Quality of service | Bandwidth-limit and minimum-bandwidth rules on ports and networks |
| BGP | Early Access |
| Firewall and VPN | Integrates with existing, compatible firewall and VPN components to offer Firewall as a Service and VPN as a Service |
| Host redundancy | Bonded network interfaces (LACP or active-backup) are recommended for production |
Sources: [Networking Overview](/private-cloud-director/virtualized-networking/networking-overview.md), [Physical Network](/private-cloud-director/virtualized-networking/physical-network.md), [Layer 2 Networking](/private-cloud-director/virtualized-networking/layer-2-networking.md), [Attach Multiple Ports from the Same Physical Network](/private-cloud-director/virtualized-networking/multi-subnet-vm-networking.md), [Add Multiple Dual Stack Ports to a VM](/private-cloud-director/tutorials/create-dual-stack-port.md), [Security Group](/private-cloud-director/virtualized-networking/security-group.md), [Load Balancer as a Service](/private-cloud-director/virtualized-networking/load-balancer.md), [DNS as a Service](/private-cloud-director/virtualized-networking/dns-as-a-service-dnsaas.md), [Networking QoS](/private-cloud-director/virtualized-networking/networking-quality-of-service-qos-configuration.md), [BGP Support](/private-cloud-director/virtualized-networking/bgp-support.md), [Architecture Overview](/private-cloud-director/introduction/architecture-overview.md), [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md).
## Storage and Data Protection
VM disks sit on the customer's existing enterprise storage through certified block storage drivers. Block storage volumes are the recommended option for production VMs.
| Area | Support |
| -------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Certified block storage backends | NetApp ONTAP (iSCSI, Fibre Channel, NFS); HPE 3PAR, Primera, Alletra and Nimble (Fibre Channel, iSCSI); Pure Storage FlashArray (iSCSI); Tintri VMstore (NFS); Infinidat InfiniBox (iSCSI, Fibre Channel); Hitachi VSP (iSCSI, Fibre Channel) |
| Supported, not certified | Dell EMC Unity, SC Series, PowerStore and PowerMax; validate in the target environment before production use |
| Other backends | A driver must be certified by Platform9 before it can be selected as a volume backend; Platform9 Support confirms certification status on request |
| Ephemeral storage | Local disk or NFS shared storage for VM root disks |
| Volume operations | Volume snapshots; extend an in-use volume; retype and migrate a volume across backends |
| Multipath | Recommended for production deployments on Fibre Channel or iSCSI |
| Volume encryption | Encrypts volumes, including boot volumes; keys are held in the built-in key manager or in HashiCorp Vault as an external key manager |
| Backup integrations | Rubrik, Commvault, Veeam, Storware and Cohesity |
| Tested limits | 10,000 volumes and 6 storage backends per region |
Sources: [Certified Block Storage Drivers and Configurations](/private-cloud-director/storage/block-storage/volume-backend-configuration-examples.md), [Supported Storage Drivers](/private-cloud-director/storage/block-storage/supported-storage-drivers.md), [Ephemeral Storage](/private-cloud-director/storage/ephemeral-storage.md), [Volume Snapshot](/private-cloud-director/storage/volume/volume-snapshot.md), [Resize a Volume](/private-cloud-director/tutorials/resize-a-volume.md), [Storage Live Migration](/private-cloud-director/storage/storage-live-migration.md), [Volume Encryption with an External Key Manager](/private-cloud-director/storage/volume/volume-encryption-external-key-manager.md), [Rubrik](/private-cloud-director/integrations/rubrik-integration-with-pcd.md), [Commvault](/private-cloud-director/integrations/commvault-integration-with-pcd.md), [Veeam](/private-cloud-director/integrations/veeam-integration-with-pcd.md), [Storware](/private-cloud-director/integrations/storware-integration-with-pcd.md), [Cohesity](/private-cloud-director/integrations/cohesity-dataprotect-with-pcd.md), [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md).
## Security and Compliance
space.vars.product\_acronym secures the platform layer: the hosts, the management plane and the isolation between tenants.
| Area | Support |
| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Workload isolation | Each workload runs as a KVM virtual machine. Every resource belongs to exactly one tenant within a domain. Separate virtualized clusters and host aggregates dedicate hosts to specific workloads |
| Access control | Role-based access control per tenant with three built-in roles: Admin, Self-Service User and ReadOnly. A six-role Granular RBAC model is in beta |
| Authentication | Enterprise SSO with Microsoft Entra ID, Okta, PingOne or Duo; multi-factor authentication with time-based one-time passwords; application credentials for API clients |
| Audit logging | Audit logs for the Identity, Compute and Networking services, retrievable from the UI |
| Encryption in transit | Hosts connect to a SaaS management plane outbound on port 443. VM console sessions are encrypted with TLS and authenticated with a per-session password. A Self-Hosted management plane accepts customer-supplied TLS certificates |
| Encryption at rest | Volume encryption, including boot volumes, with keys in the built-in key manager or HashiCorp Vault; vTPM 2.0 and Secure Boot for guests |
| Host hardening | Platform9 OS ships hardened to the CIS RHEL 10 Benchmark, Level 1 Server profile, and each host reports its live compliance status |
| Vendor remote access | Platform9 support cannot interactively access hosts by default. Advanced Remote Support is off by default and is enabled by the customer per host |
| Data residency | Workloads always run in the customer's data center on the customer's hardware. A Self-Hosted management plane, including an air-gapped one, also stays on premises |
Sources: [Architecture Overview](/private-cloud-director/introduction/architecture-overview.md), [RBAC Roles and Permissions](/private-cloud-director/identity-and-multi-tenancy/rbac-roles-and-permissions.md), [Enterprise SSO](/private-cloud-director/identity-and-multi-tenancy/enterprise-sso.md), [Login with Multi-factor Authentication](/private-cloud-director/identity-and-multi-tenancy/login-with-multi-factor-authentication.md), [Application Credentials](/private-cloud-director/identity-and-multi-tenancy/application-credentials.md), [Audit Logging](/private-cloud-director/monitoring-and-observability/audit-logging.md), [Pre-requisites](/private-cloud-director/getting-started/pre-requisites.md), [Using Custom Certificates](/private-cloud-director/getting-started/self-hosted/using-custom-certificates.md), [Volume Encryption with an External Key Manager](/private-cloud-director/storage/volume/volume-encryption-external-key-manager.md), [Virtual TPM](/private-cloud-director/virtualized-clusters/virtual-tpm.md), [Platform9 OS CIS Hardening](/private-cloud-director/getting-started/platform9-os/platform9-os-cis-hardening.md), [Advanced Remote Support](/private-cloud-director/virtualized-clusters/add-hosts-virtualized-cluster/enable-advanced-remote-support.md), [Air Gapped Installation](/private-cloud-director/getting-started/self-hosted/self-hosted-airgap-install.md).
## Operations, Lifecycle and Support
In the SaaS model Platform9 patches and upgrades the management plane, and the customer operates its hosts and the network path to them. In the Self-Hosted model the customer operates both.
| Area | Support |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Deployment models | SaaS management plane hosted and operated by Platform9; Self-Hosted management plane installed and lifecycle-managed with the `airctl` tool, including air-gapped installation |
| Upgrades | The management plane upgrades first, then the hosts. Each host is drained with maintenance mode before it upgrades |
| Monitoring | Built-in dashboards based on Prometheus and Grafana, metrics export to an external observability stack, a per-host Service Health Dashboard and alerting |
| Support tooling | Support bundle generation; Advanced Remote Support for Platform9 engineers when the customer enables it |
| Automation | The `pcdctl` CLI, the API and a Terraform provider |
| Tested scale per region | 591 active hypervisors, 9,673 powered-on VMs and 8 clusters, with up to 100 hypervisors per cluster |
Sources: [Architecture Overview](/private-cloud-director/introduction/architecture-overview.md), [Host Upgrade Runbook](/private-cloud-director/upgrade/host-upgrade-runbook.md), [Monitoring](/private-cloud-director/monitoring-and-observability/monitoring.md), [Service Health Dashboard](/private-cloud-director/monitoring-and-observability/service-health-dashboard.md), [Generate a Support Bundle](/private-cloud-director/virtualized-clusters/troubleshooting-and-log-files/generate-a-support-bundle.md), [Advanced Remote Support](/private-cloud-director/virtualized-clusters/add-hosts-virtualized-cluster/enable-advanced-remote-support.md), [pcdctl CLI](/private-cloud-director/automation-and-cli/pcdctl-command-line.md), [Terraform Provider](/private-cloud-director/automation-and-cli/terraform-provider.md), [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md).
## Related Pages
* [Architecture Overview](/private-cloud-director/introduction/architecture-overview.md): how the management plane, hosts, and services fit together
* [Configuration Maximums](/private-cloud-director/getting-started/configuration-maximums.md): tested limits per region, cluster, and VM
* [Host Operating System Support Matrix](/private-cloud-director/getting-started/host-operating-system-support-matrix.md): supported host OS per release
* [Transition from VMware](/private-cloud-director/getting-started/transition-from-vmware.md): VMware concepts and their equivalents
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://docs.platform9.com/private-cloud-director/introduction/technical-data-sheet.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.