vinzenz/hyperhive
vinzenz/hyperhive
Watch
0
0
Fork
You've already forked hyperhive
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

hivectl: batch QueueNodes polling by id set, drop remaining dag wording

Browse source
This commit is contained in:
damocles 2026-08-03 20:17:16 +02:00 • committed by mara
commit 0c4d56a585
8 changed files with 182 additions and 125 deletions
Download patch file Download diff file Expand all files Collapse all files

42
hive-c0re/src/job_queue/mod.rs
View file

@ -379,30 +379,34 @@ impl JobQueue {
.collect()
}
/// A node plus its live subtree, as generic wire nodes — the
/// `QueueNodes` polling surface behind `hivectl`'s wait/progress loop.
/// Sibling of [`Self::snapshot`] (which serves the same graph through
/// the typed `DagView`/`NodeView` projection for the dashboard's
/// `/api/state.rebuild_queue`), this one goes through
/// One or more nodes plus their live subtrees, as generic wire nodes —
/// the `QueueNodes` polling surface behind `hivectl`'s wait/progress
/// loop. Sibling of [`Self::snapshot`] (which serves the same graph
/// through the typed `DagView`/`NodeView` projection for the
/// dashboard's `/api/state.rebuild_queue`), this one goes through
/// [`GraphWire::wire_snapshot`] instead — no `Done`-node filtering, no
/// roll-up field (a node's own `state` answers that, see
/// `hive_jobq_wire`'s doc comment). Looks the id up by identity alone —
/// no assumption that it names a DAG container or a root; "just show
/// whatever the backend sends" for whatever id the caller asks about.
/// `hive_jobq_wire`'s doc comment). Looks each id up by identity
/// alone — no assumption that it names a DAG container or a root;
/// "just show whatever the backend sends" for whatever ids the caller
/// asks about. Multiple ids in one call is the normal shape for a
/// batch op (e.g. restarting every agent submits one root per agent) —
/// callers should request the whole batch together rather than poll
/// one id per round-trip.
///
/// Empty when `id` names no node in the graph. Today that only happens
/// for a genuinely unknown id: nothing prunes the graph yet
/// (bounded-prune is a Stage-C follow-up, see [`visible_dags`]), so a
/// *completed* DAG's nodes keep riding here with a terminal `state`
/// rather than disappearing — callers watching for "done" should read
/// the root's `state`, not emptiness.
/// An id with no matching node in the graph is silently dropped from
/// the result rather than erroring the whole batch — some ids in a
/// batch may already be evicted while others are still live. Today
/// that only happens for a genuinely unknown id: nothing prunes the
/// graph yet (bounded-prune is a Stage-C follow-up, see
/// [`visible_dags`]), so a *completed* DAG's nodes keep riding here
/// with a terminal `state` rather than disappearing — callers
/// watching for "done" should read the root's `state`, not absence.
#[must_use]
pub fn node_subtree(&self, id: u64) -> Vec<GraphNode> {
pub fn node_subtrees(&self, ids: &[u64]) -> Vec<GraphNode> {
let inner = self.lock();
let Some(node) = find_node(&inner, id) else {
return Vec::new();
};
inner.graph().wire_snapshot([node])
let roots: Vec<NodeId> = ids.iter().filter_map(|id| find_node(&inner, *id)).collect();
inner.graph().wire_snapshot(roots)
}
}

4
hive-c0re/src/server.rs
View file

@ -160,8 +160,8 @@ async fn dispatch(req: &HostRequest, coord: Arc<Coordinator>) -> HostResponse {
.collect();
HostResponse::dags(dags)
}
HostRequest::QueueNodes { id } => {
HostResponse::nodes(coord.job_queue.node_subtree(*id))
HostRequest::QueueNodes { ids } => {
HostResponse::nodes(coord.job_queue.node_subtrees(ids))
}
HostRequest::List => HostResponse::list(lifecycle::list().await?),
// The agents root is ours and not world-traversable, so this