RestoreTest

API group: restore.kymaros.io/v1alpha1
Kind: RestoreTest
Short name: rt
Scope: Namespaced (typically kymaros-system)

A RestoreTest defines a scheduled restore validation job. The controller reads it, restores the specified backup into an isolated sandbox namespace at each scheduled interval, runs health checks, measures RTO, and produces a RestoreReport.

Spec

Top-level fields

Field	Type	Required	Default	Description
`backupSource`	`BackupSource`	Yes	—	Identifies the backup to restore and from which provider.
`schedule`	`ScheduleConfig`	Yes	—	Controls when restore tests run.
`sandbox`	`SandboxConfig`	Yes	—	Configures the isolated sandbox namespace used during testing.
`healthChecks`	`HealthCheckRef`	No	—	References a `HealthCheckPolicy` to run after restore.
`sla`	`SLAConfig`	No	—	Defines the RTO target and alert behavior.
`notifications`	`NotificationConfig`	No	—	Configures where pass/fail notifications are sent.
`timeout`	`Duration`	No	—	Global timeout for the entire test run.
`historyLimit`	`int32`	No	`10`	Number of `RestoreReport` objects to retain. Minimum: 1.

BackupSource

Field	Type	Required	Default	Description
`provider`	`string`	Yes	—	Backup provider. Accepted values: `velero`, `kasten`, `trilio`.
`backupName`	`string`	Yes	—	Name of the backup to restore. Use `"latest"` to always select the most recent backup.
`namespaces`	`[]NamespaceMapping`	Yes	—	Source namespaces to restore. At least one entry required.
`labelSelector`	`map[string]string`	No	—	Kubernetes label selector applied when identifying the backup.

NamespaceMapping

Field	Type	Required	Default	Description
`name`	`string`	Yes	—	Source namespace name in the backup.
`sandboxName`	`string`	No	—	Override the generated sandbox namespace name for this source namespace. When omitted, the name is derived from `sandbox.namespacePrefix` and the source namespace name.

ScheduleConfig

Field	Type	Required	Default	Description
`cron`	`string`	Yes	—	Standard five-field cron expression (e.g., `"0 3 * * *"` for 03:00 daily).
`timezone`	`string`	No	`"UTC"`	IANA timezone name for cron evaluation (e.g., `"Europe/Paris"`).

SandboxConfig

Field	Type	Required	Default	Description
`namespacePrefix`	`string`	No	`"rp-test"`	Prefix for generated sandbox namespace names.
`ttl`	`Duration`	No	`"30m"`	How long the sandbox namespace lives after the test completes before automatic deletion.
`resourceQuota`	`ResourceQuotaConfig`	No	—	Resource limits applied to the sandbox namespace.
`networkIsolation`	`string`	No	`"strict"`	Network policy mode. `strict`: sandbox has no external egress. `group`: sandbox pods can reach other sandboxes in the same group but not production.

ResourceQuotaConfig

Field	Type	Required	Default	Description
`cpu`	`string`	No	—	CPU limit for the sandbox namespace (e.g., `"4"`).
`memory`	`string`	No	—	Memory limit for the sandbox namespace (e.g., `"8Gi"`).
`storage`	`string`	No	—	Total storage request limit for the sandbox namespace (e.g., `"50Gi"`).

HealthCheckRef

Field	Type	Required	Default	Description
`policyRef`	`string`	No	—	Name of a `HealthCheckPolicy` resource in the same namespace.
`timeout`	`Duration`	No	—	Maximum time to wait for all health checks defined in the referenced policy to pass.

SLAConfig

Field	Type	Required	Default	Description
`maxRTO`	`Duration`	Yes (if `sla` set)	—	Maximum acceptable restore time. The measured duration is compared against this value and recorded in the report as `rto.withinSLA`.
`alertOnExceed`	`bool`	No	—	When `true`, a notification is sent through the configured channels when measured RTO exceeds `maxRTO`.

NotificationConfig

Field	Type	Required	Default	Description
`onFailure`	`[]NotificationChannel`	No	—	Channels to notify when the test result is `fail`.
`onSuccess`	`[]NotificationChannel`	No	—	Channels to notify when the test result is `pass`.

NotificationChannel

Field	Type	Required	Default	Description
`type`	`string`	Yes	—	Notification backend. Accepted values: `slack`, `pagerduty`, `webhook`.
`channel`	`string`	No	—	Slack channel name or ID (e.g., `"#alerts"`). Only applicable when `type` is `slack`.
`webhookSecretRef`	`string`	No	—	Name of a `Secret` in the same namespace that contains the webhook URL or API token. The secret must have a key named `url` for `webhook` type or `token` for `pagerduty`.

Status

Field	Type	Description
`phase`	`string`	Current lifecycle phase: `Idle`, `Running`, `Completed`, or `Failed`.
`lastRunAt`	`Time`	Timestamp of the most recent completed test run.
`lastScore`	`int`	Confidence score (0–100) from the most recent run.
`lastResult`	`string`	Outcome of the most recent run: `pass`, `fail`, or `partial`.
`lastReportRef`	`string`	Name of the `RestoreReport` object created by the most recent run.
`nextRunAt`	`Time`	Scheduled time for the next test run, derived from `schedule.cron`.
`sandboxNamespace`	`string`	Name of the active sandbox namespace (populated during `Running` phase).
`restoreID`	`string`	Provider-specific restore operation identifier (populated during `Running` phase).
`conditions`	`[]Condition`	Standard Kubernetes condition array reflecting reconciliation state.

Print columns exposed by kubectl get rt: Phase, Score, Result, Last Run, Age.

Examples

Simple stateless application

Runs nightly at 02:00 UTC, restores the latest Velero backup of the my-app namespace, applies a 30-minute RTO target, and notifies Slack on failure.

apiVersion: restore.kymaros.io/v1alpha1
kind: RestoreTest
metadata:
  name: my-app-nightly
  namespace: kymaros-system
spec:
  backupSource:
    provider: velero
    backupName: latest
    namespaces:
      - name: my-app
  schedule:
    cron: "0 2 * * *"
    timezone: UTC
  sandbox:
    namespacePrefix: rp-test
    ttl: 30m
    networkIsolation: strict
  sla:
    maxRTO: "30m"
    alertOnExceed: true
  notifications:
    onFailure:
      - type: slack
        channel: "#ops-alerts"
        webhookSecretRef: slack-kymaros-webhook
  historyLimit: 10

Stateful application with database health checks

Restores the orders namespace using Velero, runs a custom HealthCheckPolicy that verifies the database pod and HTTP readiness endpoint, enforces a 15-minute RTO, and sends both success and failure notifications.

apiVersion: restore.kymaros.io/v1alpha1
kind: RestoreTest
metadata:
  name: orders-db-validation
  namespace: kymaros-system
spec:
  backupSource:
    provider: velero
    backupName: orders-daily
    namespaces:
      - name: orders
  schedule:
    cron: "0 3 * * *"
    timezone: "Europe/Paris"
  sandbox:
    namespacePrefix: rp-test
    ttl: 45m
    resourceQuota:
      cpu: "4"
      memory: 8Gi
      storage: 50Gi
    networkIsolation: strict
  healthChecks:
    policyRef: orders-health-policy
    timeout: 10m
  sla:
    maxRTO: "15m"
    alertOnExceed: true
  notifications:
    onFailure:
      - type: pagerduty
        webhookSecretRef: pagerduty-kymaros-token
      - type: slack
        channel: "#platform-alerts"
        webhookSecretRef: slack-kymaros-webhook
    onSuccess:
      - type: slack
        channel: "#platform-ops"
        webhookSecretRef: slack-kymaros-webhook
  timeout: 1h
  historyLimit: 30

Multi-namespace application

Restores three namespaces that together form a single application (frontend, backend API, and shared infrastructure). Each namespace maps to a distinct sandbox name to avoid collisions.

apiVersion: restore.kymaros.io/v1alpha1
kind: RestoreTest
metadata:
  name: platform-full-stack
  namespace: kymaros-system
spec:
  backupSource:
    provider: velero
    backupName: latest
    namespaces:
      - name: platform-frontend
        sandboxName: rp-frontend
      - name: platform-api
        sandboxName: rp-api
      - name: platform-infra
        sandboxName: rp-infra
    labelSelector:
      app.kubernetes.io/part-of: platform
  schedule:
    cron: "0 4 * * 0"
    timezone: "America/New_York"
  sandbox:
    namespacePrefix: rp-test
    ttl: 1h
    resourceQuota:
      cpu: "8"
      memory: 16Gi
      storage: 100Gi
    networkIsolation: group
  healthChecks:
    policyRef: platform-health-policy
    timeout: 15m
  sla:
    maxRTO: "45m"
    alertOnExceed: true
  notifications:
    onFailure:
      - type: webhook
        webhookSecretRef: incident-webhook
  timeout: 2h
  historyLimit: 20

kubectl quick reference

# List all RestoreTest resources
kubectl get rt -n kymaros-system

# Describe a specific test
kubectl describe rt my-app-nightly -n kymaros-system

# Watch status in real time
kubectl get rt -n kymaros-system -w

# Edit a test
kubectl edit rt my-app-nightly -n kymaros-system

Spec​

Top-level fields​

BackupSource​

NamespaceMapping​

ScheduleConfig​

SandboxConfig​

ResourceQuotaConfig​

HealthCheckRef​

SLAConfig​

NotificationConfig​

NotificationChannel​

Status​

Examples​

Simple stateless application​

Stateful application with database health checks​

Multi-namespace application​

kubectl quick reference​