Management

Tenant Management #

Defines the lifecycle and operational model for tenant workloads.


Purpose #

Tenant management provides:

  • Lifecycle control — Create, update, and destroy tenant environments
  • Observability — Logs, metrics, and alerting scoped to tenants
  • Access control — Who can manage which tenants
  • Operational clarity — Clear boundaries between tenants

Tenant Lifecycle #

Create #

Creating a new tenant involves:

  1. Reserve VLAN and subnet — Allocate from tenant IP range
  2. Configure network infrastructure — Add VLAN interface, DHCP scope, firewall zone
  3. Provision tenant — Deploy VMs and DNS records via Terraform
  4. Configure observability — Set up log/metric collection for tenant

Update #

Updating a tenant may include:

  • Adding or removing VMs
  • Changing resource allocations
  • Updating firewall rules
  • Modifying DNS records

Updates are applied via Terraform for tenant resources, automation for network configuration.

Destroy #

Destroying a tenant:

  1. Destroy tenant resources — Terraform destroys VMs and DNS records
  2. Remove network config — Delete VLAN, DHCP scope, firewall zone
  3. Archive data — Backup logs and metrics if required
  4. Release resources — Return VLAN ID and subnet to pool

Tenant Observability #

Logs #

Tenant logs are:

  • Collected by management plane observability stack
  • Tagged with tenant identifier
  • Queryable by tenant scope
  • Retained per tenant policy

Metrics #

Tenant metrics include:

  • VM resource utilization (CPU, memory, disk, network)
  • Application-level metrics (if instrumented)
  • Network traffic volumes

Alerting #

Alerts may be configured:

  • Per-tenant thresholds
  • Tenant-specific notification channels
  • Escalation policies

Access Control #

Tenant Boundaries #

Each tenant is an isolated security domain:

  • No cross-tenant network access by default
  • Separate credentials and access paths
  • Independent lifecycle management

Administrative Access #

RoleAccess
Platform adminAll tenants, substrate infrastructure
Tenant adminSpecific tenant(s), scoped access

Access is controlled via:

  • SSH key distribution
  • Jump host access policies
  • Firewall rules

Relationship to Substrate Management #

Tenant management is distinct from substrate management:

AspectSubstrate ManagementTenant Management
ScopeInfrastructure (router, hypervisors)Workloads (VMs, applications)
ToolingAutomation-firstTerraform-first
LifecycleRare changes, high stabilityFrequent changes, agile
AuthorityPlatform admins onlyMay delegate to tenant admins

The substrate Management Plane provides services that tenants consume (DNS, DHCP, observability).


Tenant Isolation Principles #

Blast Radius Containment #

A problem in one tenant should not affect others:

  • Network isolation via VLANs
  • Resource quotas (future)
  • Independent lifecycle

No Shared State #

Tenants do not share:

  • Databases
  • File storage
  • Credentials
  • Configuration

Shared services (DNS, NAT) are substrate-level, not tenant-level.

Explicit Dependencies #

If a tenant depends on another service:

  • Document the dependency
  • Create explicit firewall rules
  • Monitor the dependency path

Operational Runbooks #

Common tenant operations:

OperationRunbook
Create new tenantReserve VLAN, configure router, deploy VMs
Add VM to tenantUpdate Terraform, apply, verify
Debug tenant networkCheck VLAN, DHCP, firewall rules
Investigate tenant issueQuery tenant-scoped logs and metrics
Decommission tenantDestroy VMs, clean up network, archive data

Summary #

  1. Tenants have explicit lifecycle: create, update, destroy
  2. Observability (logs, metrics, alerts) is scoped per tenant
  3. Access control separates platform admins from tenant admins
  4. Tenant management is distinct from substrate management
  5. Isolation principles prevent cross-tenant impact
Page last modified: March 8, 2026