generated: '2026-07-19'
method: searched
source: https://docs.kubeark.com/
note: >-
  Kubeark ships no public HTTP API and no OpenAPI, so the usual request/response
  conventions (pagination params, error envelope, rate-limit headers, request-id header)
  are NOT published and are recorded as unknown rather than guessed. What IS documented is
  the execution semantics of the workflow engine, captured below. NO Idempotency pointer
  is emitted: Kubeark documents no idempotency key, and its run modes are concurrency
  controls, not idempotent-replay semantics.
authentication:
  style: vault-backed API tokens and kubeconfig for workflow nodes; SSO/OIDC/SAML2/LDAP
    for platform access
  artifact: authentication/kubeark-authentication.yml
idempotency:
  supported: false
  header: null
  note: >-
    No idempotency key, no request-replay contract and no deduplication window is
    documented anywhere in the Kubeark docs.
concurrency:
  documented: true
  mechanism: workflow run modes
  since: '2024-10'
  modes:
  - id: parallel
    description: Executions run in parallel (the original behavior).
  - id: terminate
    description: Each new execution cancels any previously running executions.
  - id: queue
    description: New executions wait in a queue until previous ones finish.
  configured_in: workflow Settings modal
timeouts:
  documented: true
  default: 24 hours
  configurable: true
  note: Users can set a custom timeout per workflow; the default is 24 hours.
pagination:
  documented: false
  note: >-
    Pagination exists in the UI (execution history gained pagination in the October 2024
    release) but no API pagination parameters or response fields are published.
error_handling:
  envelope: unknown
  node_level: true
  mechanism: >-
    Each workflow node carries an On Error setting controlling behavior when the node
    fails; a Condition field allows conditional execution. No API error envelope or error
    code registry is published.
  artifact: null
versioning:
  scheme: major product version of the self-hosted platform
  api_versioning: none published
  artifact: lifecycle/kubeark-lifecycle.yml
rate_limiting:
  documented: false
  note: >-
    No rate-limit headers or quotas are published. The closest documented control is a
    prepaid workspace option that limits total execution time per workspace.
observability:
  audit_logs: true
  audit_docs: https://docs.kubeark.com/logs-global
  filters:
  - username
  - event (Added / Updated / Deleted)
  - model (Workspace, Workflow, Trigger, Source, Cron Job)
  - date range
  execution_history: https://docs.kubeark.com/execution-history
  metrics:
    status: preview
    since: '2024-10'
    description: Workflow metrics providing insights into workflow performance and
      execution data.
secrets:
  mechanism: Local Vault (workspace scope) and Global Vault (platform scope)
  note: Node inputs reference vault entries rather than embedding secrets in the workflow
    definition.
variables:
  local: Created on each script node; values can be set or prompted at the start of the
    node.
  global: Predefined variables usable throughout the workflow; cannot be created inside
    individual nodes.
event_surface:
  artifact: asyncapi/kubeark-webhooks.yml