ArchiveExecutionOutput

"""
One run of an archive job. Written exclusively by the archive worker and scheduler; immutable from the GraphQL API. Carries the run's outcome plus enough context (schema snapshot, captured script logs, error message) to debug a failed script.
"""
type ArchiveExecutionOutput {
  """TypeID of this execution, prefixed "arx" (e.g. "arx_01K…")."""
  id: ArchiveExecutionID!
  """
  TypeID of the owning tenant, prefixed "ten". Always the caller's tenant.
  """
  tenantId: TenantID!
  """
  TypeID of the archive job this run belongs to, prefixed "arj" (e.g. "arj_01K…").
  """
  archiveJobId: ArchiveJobID!
  """
  Current lifecycle state of the run; PENDING and RUNNING are the only non-terminal values.
  """
  status: ArchiveExecutionStatus!
  """
  ISO-8601 timestamp. Set when the row is created (enqueued), then overwritten with the wall-clock time the worker actually began executing the script.
  """
  startedAt: String!
  """
  ISO-8601 timestamp at which the run reached a terminal status; null while PENDING or RUNNING.
  """
  finishedAt: String
  """
  Number of rows the script emitted. 0 when the run drained no rows (still SUCCEEDED, but nothing is uploaded). Null until the run finishes, and for runs that never produced a result (SKIPPED, CANCELLED, failed-before-drain).
  """
  rowCount: Int
  """
  Size in bytes of the object uploaded to the storage backend. 0 for a zero-row run. Null until the upload completes, and for runs that never uploaded.
  """
  bytesWritten: Int
  """
  Rendered storage key (path) of the uploaded archive file on the job's storage backend. Present only when the run SUCCEEDED and uploaded at least one row; null for zero-row, skipped, cancelled, and failed-before-upload runs.
  """
  outputKey: String
  """
  JSON array describing the Arrow schema the script declared for this run, as returned by its Lua exports.schema() hook: [{ name, type, nullable }]. Records what columns the archived file actually holds. Null if the script failed before declaring a schema.
  """
  schemaSnapshot: String
  """
  User-facing failure reason for a FAILED run (e.g. the Lua script error). Also carries the reason a SKIPPED run was never started. Null otherwise.
  """
  errorMessage: String
  """
  JSON array of print()/warn() lines the tenant's Lua script emitted during this run: [{ level, message, ts }]. Lets users debug scripts without server-log access. Null when the script never ran.
  """
  logCapture: String
  """
  JSON-encoded value the script returned from its exports.checkpoint() hook at end-of-run; shape is entirely script-defined (a cursor, a timestamp, a table). Persisted only for SUCCEEDED runs — a failed run leaves the previous success's checkpoint in place so the next attempt resumes from there. Null if the script defines no checkpoint hook.
  """
  checkpoint: String
  """
  What caused this run to be enqueued: the scheduler, or a named user via runArchiveJob.
  """
  trigger: ArchiveExecutionTriggerOutput!
  """
  ISO-8601 timestamp at which the execution row was inserted, i.e. when the run was enqueued.
  """
  createdAt: String!
  """
  ISO-8601 timestamp of the last write to the row — effectively the last status transition, since executions are never edited through the API.
  """
  updatedAt: String!
}