Get approval info for a suspended flow/WAC job. Returns form schema, approval rules, and whether the current user can approve. Either a valid token query parameter or an authenticated session is required.
One failed run is saved per script per execution principal, so a retry resumes that one — not necessarily the run being looked at. Answers about THIS run alone: its own id when a retry submitted by this caller would resume it, null otherwise (a later run holds the state, the run left nothing to rebuild, or the caller’s runs execute as another principal). Authorized through the job, like run_progress.
A retry requires dbt_retry_job, and the run form has no job id to hand. Returns the one the caller’s own retry would land on — the saved failure for the principal the run would execute as, which is the script’s author for an on_behalf_of script and the caller otherwise — and only when that run is one the caller may read. Null when there is nothing to resume, or nothing this caller may be told about.
The workspace asset graph with its dbt half pinned to what a single run built. Only a dynamic descriptor (a {{ }} placeholder in vars) leaves a per-run snapshot; without one for this job the pinned version’s own graph answers, so a run page may use this route unconditionally. Authorized through the job, the same gate as run_progress — pinning to a run is job-scoped, which is why assets/graph cannot do it.
Returns a stateless {job_id}.{hmac} token that grants anyone holding it — including logged-out visitors, who land on the minimal public run page — read access to this job (and its flow subtree) via a view_token query param or X-View-Token header. Only callable by a user who can already read the job.
Returns a stateless {job_id}.{hmac} token that grants an authenticated workspace member read access to this job (and its flow subtree) via a view_token query param or X-View-Token header. Only callable by a user who can already read the job.
Live materialization state for the relations a single job writes, keyed by relation. Polled while a run is in flight so a graph can move a node as its model completes. Authorized through the job, so a caller who may not read it is refused rather than told the run has no progress. An empty list means the job exists and has recorded nothing yet, or is unknown to this workspace.
Returns the dispatched asset-trigger edges (producer job → child job) whose subscriber lives under path_start. Lets a pipeline view reconstruct the cascade tree of a folder by job id and group connected runs. Visibility follows the producer job’s RLS.
Returns the chronological log of decisions the asset-trigger dispatcher made after this producer job completed. Each row is one (subscriber, asset write) decision: dispatched (with child_job_id), join_pending (with received_inputs / required_inputs / partition), or skipped (with reason). Rows are reaped automatically when the producer’s v2_job row is deleted by the retention sweep.
Marks failed jobs as handled so triage surfaces stop showing them as failures. The job status itself is unchanged. Returns the ids actually resolved: ids that are not visible to the caller or did not fail are silently skipped.
Resume or cancel a suspended flow/WAC job. Uses approval rules to determine authorization. Either a valid approval_token or an authenticated session is required.