The pool is now rendered from packages/apps/tenant/templates/cilium-lb-pool.yaml instead of packages/extra/ingress/templates/cilium-lb-pool.yaml. Cilium LB IPAM forbids overlapping CIDRs across pools regardless of serviceSelector, so both the ingress-loadBalancer path and the upcoming per-tenant Gateway in #2470 cannot each own their own pool on the same publishing.externalIPs range. The tenant chart is the natural per-tenant owner — it already creates the Namespace, the cozystack-values Secret, and the HelmReleases for both ingress and gateway. The new pool uses a namespace-only serviceSelector (io.kubernetes.service.namespace: <ns>), which matches any LoadBalancer Service in the tenant namespace. The metadata.name changed from <trim>-ingress to <trim>-exposure to reflect that the pool is not ingress-specific. Only the ingress-loadBalancer signal is wired in this commit (_cluster.expose-mode=loadBalancer plus .Values.ingress=true on the publishing tenant). The gateway branch is added in #2470 on top of this commit — it rebases, drops its own packages/extra/gateway/templates/cilium-lb-pool.yaml, and adds an OR branch for .Values.gateway in the tenant template. Pool-rendering unit tests moved from packages/extra/ingress/tests/ to packages/apps/tenant/tests/. The ingress chart tests keep the Service-level asserts. packages/apps/tenant/Makefile gains a test target so hack/helm-unit-tests.sh picks up the new suite. Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Aleksei Sviridkin <f@lex.la> |
||
|---|---|---|
| .. | ||
| charts | ||
| logos | ||
| templates | ||
| tests | ||
| .helmignore | ||
| Chart.yaml | ||
| Makefile | ||
| README.md | ||
| values.schema.json | ||
| values.yaml | ||
Tenant
A tenant is the main unit of security on the platform. The closest analogy would be Linux kernel namespaces.
Tenants can be created recursively and are subject to the following rules:
Tenant naming
Tenant names must be alphanumeric:
- Lowercase letters (
a-z) and digits (0-9) only - Must start with a lowercase letter
- Dashes (
-) are not allowed, unlike with other services - Maximum length depends on the cluster configuration (Helm release prefix and root domain)
This restriction exists to keep consistent naming in tenants, nested tenants, and services deployed in them.
A tenant cannot be named foo-bar because parsing internal resource names like tenant-foo-bar would be ambiguous.
For example:
- The root tenant is named
root, but internally it's referenced astenant-root. - A nested tenant could be named
foo, which would result intenant-fooin service names and URLs.
Unique domains
Each tenant has its own domain.
By default, (unless otherwise specified), it inherits the domain of its parent with a prefix of its name.
For example, if the parent had the domain example.org, then tenant-foo would get the domain foo.example.org by default.
Kubernetes clusters created in this tenant namespace would get domains like: kubernetes-cluster.foo.example.org
Example:
tenant-root (example.org)
└── tenant-foo (foo.example.org)
└── kubernetes-cluster1 (kubernetes-cluster1.foo.example.org)
Nesting tenants and reusing parent services
Tenants can be nested. A tenant administrator can create nested tenants using the "Tenant" application from the catalogue.
Higher-level tenants can view and manage the applications of all their children tenants. If a tenant does not run their own cluster services, it can access ones of its parent.
For example, you create:
- Tenant
tenant-u1with a set of services likeetcd,ingress,monitoring. - Tenant
tenant-u2nested intenant-u1.
Let's see what will happen when you run Kubernetes and Postgres under tenant-u2 namespace.
Since tenant-u2 does not have its own cluster services like etcd, ingress, and monitoring,
the applications running in tenant-u2 will use the cluster services of the parent tenant.
This in turn means:
- The Kubernetes cluster data will be stored in
etcdfortenant-u1. - Access to the cluster will be through the common
ingressoftenant-u1. - Essentially, all metrics will be collected in the
monitoringfromtenant-u1, and only that tenant will have access to them.
Example:
tenant-u1
├── etcd
├── ingress
├── monitoring
└── tenant-u2
├── kubernetes-cluster1
└── postgres-db1
Parameters
Common parameters
| Name | Description | Type | Value |
|---|---|---|---|
host |
The hostname used to access tenant services (defaults to using the tenant name as a subdomain for its parent tenant host). | string |
"" |
etcd |
Deploy own Etcd cluster. | bool |
false |
monitoring |
Deploy own Monitoring Stack. | bool |
false |
ingress |
Deploy own Ingress Controller. | bool |
false |
seaweedfs |
Deploy own SeaweedFS. | bool |
false |
schedulingClass |
The name of a SchedulingClass CR to apply scheduling constraints for this tenant's workloads. | string |
"" |
resourceQuotas |
Define resource quotas for the tenant. | map[string]quantity |
{} |
Configuration
Resource Quotas
The resourceQuotas parameter allows you to limit resources available to the tenant. Supported keys include:
Compute resources (converted to requests.X and limits.X):
cpu- Total CPU cores (e.g.,"4"or"500m")memory- Total memory (e.g.,"4Gi"or"512Mi")ephemeral-storage- Ephemeral storage limit (e.g.,"10Gi")storage- Total persistent storage (e.g.,"100Gi")
Object count quotas (passed as-is):
pods- Maximum number of podsservices- Maximum number of servicesservices.loadbalancers- Maximum number of LoadBalancer servicesservices.nodeports- Maximum number of NodePort servicesconfigmaps- Maximum number of ConfigMapssecrets- Maximum number of Secretspersistentvolumeclaims- Maximum number of PVCs
Example:
resourceQuotas:
cpu: 4
memory: 4Gi
storage: 10Gi
services.loadbalancers: "3"
pods: "50"