Pipeline health table
Joins pipeline structure with the latest build log into a single
per-node health table. Failed nodes sort first so the root cause is
visible without extra calls. Columns: - name — node name
(String) - runtime — e.g. “T”, “R”, “Python” (String) -
status — build status from the latest matching log
(“Completed”, “Errored”, “SoftFailed”, “Cached”, “Skipped”, …), or NA
when the pipeline has no matching build log yet - duration
— node build duration in seconds (Float), or NA - path —
Nix store path of the node output, or NA - error —
code: message for failed nodes (truncated), or NA
Pipeline): The pipeline to
summarize.One row per node, failed nodes first.
pipeline_status(p)
pipeline_status(p) |> filter($status == "Errored")