Get started

System requirements

What Telark needs from your cluster, what the chart installs, and how much capacity to reserve.

Cluster

RequirementDetail
Kubernetes1.30 or newer, enforced by the chart's kubeVersion, because the CRD write guard uses ValidatingAdmissionPolicy, GA in 1.30. 1.33+ is the tested target.
APIsOnly GA APIs: apps/v1, autoscaling/v2, networking.k8s.io/v1, policy/v1, apiextensions.k8s.io/v1, admissionregistration.k8s.io/v1.
DistributionsEKS, GKE, AKS, OpenShift, and upstream Kubernetes.
Helm3.x.
Install permissionscluster-admin. The chart registers CRDs, ClusterRoles, ValidatingAdmissionPolicies and, by default, Kyverno's admission webhooks. Afterwards Telark runs under its own service accounts.
Admission webhooksMust be allowed. Some hardened distributions restrict them.
NetworkPolicyOptional. The chart's ingress policies (on by default) only take effect with a CNI that enforces them, such as Calico or Cilium.

Storage

The exporter mounts two PersistentVolumeClaims with the same class: one for snapshots and one for protection plan reports. By default they are ReadWriteOnce claims from the cluster's default StorageClass, which managed clusters and kind, minikube or k3d already have. Two exporter replicas need a ReadWriteMany class instead: see Choose storage.

ModeSnapshotsReports
standard (default)512 MiB512 MiB
performance2 GiB2 GiB
minimal512 MiB512 MiB

Size the reports volume for the reports you keep. Reports kept after their plan is deleted are never swept, and each repeating plan keeps its newest 100 window reports of up to 16 MiB each, so a busy Enforce series can hold about 1.6 GiB.

The bundled Redis adds a 2 GiB volume, NATS a 1 GiB volume, and the model runtime a 6 GiB volume.

Volume expansion is recommended: Settings → Storage shows how full the snapshot and report volumes are.

Capacity

app.mode sizes every Telark service from one flag.

ModeReplicas per serviceRequests per podLimits per podFor
minimal1, fixed50m CPU, 64 MiB500m, 256 MiBEvaluation and dev. Fits one 2 vCPU / 8 GiB node.
standard (default)1, scales to 3 on CPU100m CPU, 128 MiB1 CPU, 512 MiBSmall and mid-size production.
performance1, scales to 5 on CPU250m CPU, 512 MiB2 CPU, 2 GiBLarge clusters or a high change rate. Disruption budgets on.

The exporter never autoscales: it runs services.exporter.replicas, one by default and always one in minimal.

Every Telark service container also requests 50 MiB of ephemeral storage and may use up to 2 GiB, the same in every mode.

By default the chart also installs these dependencies, with defaults that do not change with the mode:

DependencyWhyOpt out
Kyverno v1.19.2 (chart 3.9.2, two admission replicas)Enforces protection plans at admissionkyverno.enabled=false to use your own: 1.18 or later, serving the policies.kyverno.io CRDs and keeping its PolicyViolation Events
RedisLeader election, queues, caches, insightsRequired
NATS JetStreamApplication change events from discovery to the notifierRequired
metrics-serverCPU metrics for autoscalingmetrics-server.enabled=false if the cluster ships one
OllamaThe local model runtime for Insights. Requests 250m CPU and 1.5 GiB memory, 2 CPU limitollama.enabled=false

Network

The chart restricts ingress, not egress. If you add egress rules of your own, allow:

  • The Kubernetes API server and DNS from every Telark pod.
  • Traffic between Telark pods and to Redis and NATS in the release namespace.
  • The auth service to Google, only if Google sign-in fetches Google's signing keys.
  • The Ollama pod to the internet over HTTPS, only while app.ollama.autoPull=true (the default) so it can download the model. Air-gapped installs set it to false and need no egress.
  • The analyzer to your own Ollama-API endpoint, only if you set app.ollama.runtimeUrl.
  • The notifier to your webhook targets, only if you set app.webhooks.

Telark needs no managed database, no object store and no outbound connection to a Telark service.

Next