Compare commits

...
7 changed files with 222 additions and 110 deletions

2
Cargo.lock generated
View file

@ -4563,6 +4563,8 @@ dependencies = [
"async-nats",
"axum",
"futures-util",
"hive-jobq",
"hive-jobq-wire",
"serde",
"serde_json",
"swarm-queue-client",

View file

@ -580,35 +580,23 @@ fn build_approval_views(approvals: Vec<Approval>) -> Vec<ApprovalView> {
out
}
/// `/api/jobq/graph` query string. Today's only field is `states`: a
/// comma-separated allow-list of `hive_jobq::State` names (`"Pending"`,
/// `"Running"`, ...). Empty / absent ⇒ no filter (current behaviour, every
/// visible root). Set ⇒ only **root** groups whose own state is named are
/// served — a root's state is already its subtree's rolled-up answer (see
/// `hive_jobq_wire`'s doc), so filtering the root filters the whole group.
/// Unknown tokens are silently ignored (an unrecognised name matches
/// nothing rather than erroring the whole request), mirroring
/// `DashboardStreamQuery::kinds` above.
/// `/api/jobq/graph` query string — the generic jobq/wire query shape (any
/// host serving a `GraphWire` projection over HTTP takes the same one; see
/// `hive_jobq_wire::parse_states`, which does the actual parsing this side
/// of the query string). Today's only field is `states`: a comma-separated
/// allow-list of `hive_jobq::State` names (`"Pending"`, `"Running"`, ...).
/// Empty / absent ⇒ no filter (current behaviour, every visible root). Set
/// ⇒ only **root** groups whose own state is named are served — a root's
/// state is already its subtree's rolled-up answer (see `hive_jobq_wire`'s
/// doc), so filtering the root filters the whole group. Unknown tokens are
/// silently ignored (an unrecognised name matches nothing rather than
/// erroring the whole request), mirroring `DashboardStreamQuery::kinds`
/// above.
#[derive(Deserialize, Default, IntoParams)]
pub(super) struct JobqGraphQuery {
states: Option<String>,
}
/// Parses [`JobqGraphQuery::states`] into the list
/// [`crate::job_queue::JobQueue::graph_snapshot`] wants. `None` when absent
/// or when every token failed to parse — both mean "no filter" rather than
/// "match nothing", so an empty/garbled query reads as the unfiltered call
/// it replaces rather than an empty result set.
fn parse_states(raw: Option<&str>) -> Option<Vec<hive_jobq::State>> {
let states: Vec<hive_jobq::State> = raw?
.split(',')
.map(str::trim)
.filter(|s| !s.is_empty())
.filter_map(|s| serde_json::from_value(serde_json::Value::String(s.to_owned())).ok())
.collect();
(!states.is_empty()).then_some(states)
}
#[utoipa::path(
get,
path = "/api/jobq/graph",
@ -617,10 +605,14 @@ fn parse_states(raw: Option<&str>) -> Option<Vec<hive_jobq::State>> {
(status = 200, description = "every node of every retained job group, \
as generic `hive_jobq` graph nodes: identity, the parent tree, \
dependency edges with their accepted-outcome sets, lifecycle, and \
one opaque per-node payload. Group roots ride as ordinary nodes \
(`parent: null`) and `Done` nodes are not filtered by default a \
consumer renders the graph without knowing what any node means. \
`?states=` narrows to root groups in the named states.",
one opaque per-node payload. This is the generic jobq/wire shape \
(`hive_jobq_wire::GraphWire::wire_snapshot`), not a hive-c0re-only \
projection a `swarm-controller` serving its own graph responds \
with the same shape at the same path. Group roots ride as \
ordinary nodes (`parent: null`) and `Done` nodes are not \
filtered by default a consumer renders the graph without \
knowing what any node means. `?states=` narrows to root groups \
in the named states.",
body = Vec<hive_jobq_wire::GraphNode>),
),
tag = "state_snapshot"
@ -629,7 +621,7 @@ pub(super) async fn jobq_graph(
State(state): State<AppState>,
axum::extract::Query(q): axum::extract::Query<JobqGraphQuery>,
) -> axum::Json<Vec<hive_jobq_wire::GraphNode>> {
let states = parse_states(q.states.as_deref());
let states = hive_jobq_wire::parse_states(q.states.as_deref());
axum::Json(state.coord.job_queue.graph_snapshot(states.as_deref()))
}
@ -639,7 +631,9 @@ pub(super) async fn jobq_graph(
responses(
(status = 200, description = "counts by lifecycle state over the same \
groups `/api/jobq/graph` serves, as `(state, nodes, roots)` \
triples. Every state is present, zero counts included, in a fixed \
triples the generic jobq/wire roll-up shape \
(`hive_jobq_wire::state_rollup`), same as `/api/jobq/graph` \
above. Every state is present, zero counts included, in a fixed \
order a consumer renders a summary (\"3 running · 2 queued\") \
without fetching the graph and without re-deriving the tally. \
`roots` counts groups, `nodes` counts every step at any depth: one \

View file

@ -351,15 +351,16 @@ impl JobQueue {
/// still-matching descendant one level higher rather than hiding or
/// orphaning it.
///
/// The projection itself is [`hive_jobq_wire`]'s; all this layer supplies
/// is *which* nodes to show — see [`visible_roots`] for why the graph
/// can't decide the root-visibility half of that for itself.
/// The projection and state filter are both [`hive_jobq_wire`]'s; this
/// layer only supplies *which roots* are in view (see [`visible_roots`]
/// for why the graph can't decide that itself) — an older, different
/// question than state filtering, and neither replaces the other.
#[must_use]
pub fn graph_snapshot(&self, states: Option<&[State]>) -> Vec<GraphNode> {
let inner = self.lock();
let roots = visible_roots(&inner);
let nodes = inner.graph().wire_snapshot(roots);
filter_nodes_by_state(nodes, states)
hive_jobq_wire::filter_nodes_by_state(nodes, states)
}
/// Per-state counts over the **same** groups [`JobQueue::graph_snapshot`]
@ -414,26 +415,6 @@ fn find_node(sched: &Sched, id: u64) -> Option<NodeId> {
.find_map(|n| (n.id.get() == id).then_some(n.id))
}
/// [`JobQueue::graph_snapshot`]'s `states` ask, applied to the already-
/// projected node list: keeps every node — root or descendant — whose own
/// `state` is named.
///
/// Applied *after* [`GraphWire::wire_snapshot`] rather than as a root
/// pre-filter — narrowing which roots are visible at all is
/// [`visible_roots`]'s job (a different question: how much settled work is
/// retained, full stop); this is "of what's retained and live, which
/// individual nodes does the caller want shown right now." `None` (or an
/// unrecognised/absent query) is the identity filter.
fn filter_nodes_by_state(nodes: Vec<GraphNode>, states: Option<&[State]>) -> Vec<GraphNode> {
let Some(states) = states else {
return nodes;
};
nodes
.into_iter()
.filter(|n| states.contains(&n.state))
.collect()
}
/// The visible **group** set for [`JobQueue::graph_snapshot`]: every live group
/// root, plus the newest [`MAX_HISTORY_DAGS`] settled ones.
///

View file

@ -1527,60 +1527,6 @@ fn graph_snapshot_states_filters_by_node_state() {
);
}
/// One hand-built node, bypassing the scheduler entirely — `filter_nodes_by_state`
/// is a pure `Vec<GraphNode> -> Vec<GraphNode>` transform, so this exercises it
/// directly rather than trying (and failing, per this module's own rule) to
/// drive a live group into a genuinely mixed state through the real queue.
fn node(id: u64, parent: Option<u64>, state: State) -> GraphNode {
GraphNode {
id,
parent,
state,
deps: Vec::new(),
created_at: Utc::now(),
started_at: None,
finished_at: None,
error: None,
payload: hive_jobq_wire::NodePayload {
label: id.to_string(),
data: serde_json::Value::Null,
},
}
}
/// The case `graph_snapshot_states_filters_by_node_state` above can't reach:
/// a still-live group (root `Running`) holding a mix of already-`Done` and
/// still-`Pending` steps. Filtering out `Done` must drop exactly the `Done`
/// node and nothing else — the root and the `Pending` sibling both stay,
/// even though the root itself isn't in the requested state set.
#[test]
fn filter_nodes_by_state_keeps_matching_nodes_from_a_mixed_state_tree() {
let nodes = vec![
node(1, None, State::Running), // root: whole group still live
node(2, Some(1), State::Done), // finished step, should be hidden
node(3, Some(1), State::Pending), // not-yet-run step, should stay
];
let mut kept: Vec<u64> =
filter_nodes_by_state(nodes.clone(), Some(&[State::Running, State::Pending]))
.into_iter()
.map(|n| n.id)
.collect();
kept.sort_unstable();
assert_eq!(
kept,
vec![1, 3],
"Done is filtered out even though its still-live parent (root) isn't"
);
let mut unfiltered: Vec<u64> = filter_nodes_by_state(nodes, None)
.into_iter()
.map(|n| n.id)
.collect();
unfiltered.sort_unstable();
assert_eq!(unfiltered, vec![1, 2, 3], "None is the identity filter");
}
#[test]
fn error_truncation_cuts_on_a_char_boundary() {
// `truncate_error` is a pure `&str -> String`. This used to submit a DAG,

View file

@ -354,6 +354,46 @@ fn state_index(state: State) -> usize {
}
}
/// Parses a `?states=` query value (comma-separated [`State`] names, e.g.
/// `"Running,Pending"`) into the filter [`filter_nodes_by_state`] wants.
///
/// Absent or entirely-unparseable input is "no filter," never "match
/// nothing" — an empty/garbled query reads as the unfiltered call it
/// replaces rather than as a request for zero results. Unrecognised tokens
/// are silently dropped rather than erroring the whole request.
///
/// Lives here rather than in each host's own dashboard/API layer because
/// every host that serves [`GraphWire::wire_snapshot`] over HTTP wants the
/// same query shape — a second, independently-typed copy is exactly the
/// kind of drift this crate exists to prevent (see the module doc comment).
#[must_use]
pub fn parse_states(raw: Option<&str>) -> Option<Vec<State>> {
let states: Vec<State> = raw?
.split(',')
.map(str::trim)
.filter(|s| !s.is_empty())
.filter_map(|s| serde_json::from_value(serde_json::Value::String(s.to_owned())).ok())
.collect();
(!states.is_empty()).then_some(states)
}
/// Keeps only the nodes whose own `state` is named in `states` — a flat
/// filter on [`GraphNode::state`], no special-casing for a root vs. a
/// descendant (a root's own state already *is* its subtree's rolled-up
/// answer, per the module doc comment, so filtering the root filters the
/// whole group). `None` is a no-op, not "match nothing," mirroring
/// [`parse_states`]'s own absent-input rule.
#[must_use]
pub fn filter_nodes_by_state(nodes: Vec<GraphNode>, states: Option<&[State]>) -> Vec<GraphNode> {
let Some(states) = states else {
return nodes;
};
nodes
.into_iter()
.filter(|n| states.contains(&n.state))
.collect()
}
fn wire_node<N: WireNode, R: WireResource>(node: &hive_jobq::Node<N, R>) -> GraphNode {
let id = node.id.get();
GraphNode {
@ -400,7 +440,8 @@ mod tests {
use super::{
ALL_STATES, BitFlags, GraphDep, GraphNode, GraphWire, NodePayload, State, StateSchema,
TerminalState, TerminalStateSchema, WireId, WireNode, state_rollup,
TerminalState, TerminalStateSchema, WireId, WireNode, filter_nodes_by_state, parse_states,
state_rollup,
};
/// The mirrors document what the wire actually says — a mirror that
@ -630,6 +671,50 @@ mod tests {
);
}
#[test]
fn parse_states_covers_absent_malformed_and_valid() {
assert_eq!(parse_states(None), None, "no query is no filter");
assert_eq!(
parse_states(Some("")),
None,
"an empty query is no filter, not a request for zero results"
);
assert_eq!(
parse_states(Some("not-a-state,also-not-one")),
None,
"every token failing to parse is still no filter"
);
assert_eq!(
parse_states(Some(" Running , Pending ,bogus")),
Some(vec![State::Running, State::Pending]),
"trims whitespace, keeps the valid tokens, drops the unrecognised one"
);
}
#[test]
fn filter_nodes_by_state_keeps_matching_nodes_from_a_mixed_state_tree() {
let nodes = vec![
node(1, None, State::Running, Vec::new()), // root: whole group still live
node(2, Some(1), State::Done, Vec::new()), // finished step, should be hidden
node(3, Some(1), State::Pending, Vec::new()), // not-yet-run step, should stay
];
let mut kept: Vec<WireId> =
filter_nodes_by_state(nodes.clone(), Some(&[State::Running, State::Pending]))
.into_iter()
.map(|n| n.id)
.collect();
kept.sort_unstable();
assert_eq!(
kept,
vec![1, 3],
"Done is filtered out even though its still-live parent (root) isn't"
);
let unfiltered = filter_nodes_by_state(nodes, None);
assert_eq!(unfiltered.len(), 3, "no states given is a no-op, not empty");
}
#[test]
fn an_empty_payload_slot_is_omitted_from_the_wire() {
// The opaque slot costs nothing when a host has nothing to say.

View file

@ -18,6 +18,14 @@ anyhow.workspace = true
async-nats = { workspace = true, features = ["kv"] }
axum.workspace = true
futures-util.workspace = true
# The graph itself, held directly rather than behind a c0re-style wrapper
# module — that layering (`hive-c0re::job_queue`) is partially legacy (predates
# `hive-jobq`'s extraction into its own crate) and this daemon does not need it
# repeated. No `Scheduler` yet either: nothing here submits or executes a job,
# so there is nothing to schedule — just a `Graph` for the read-only endpoints
# to serve.
hive-jobq.workspace = true
hive-jobq-wire.workspace = true
serde.workspace = true
serde_json.workspace = true
# The queue connect (token mint + auth callback + reconnect) is shared with

View file

@ -23,16 +23,49 @@
use std::os::unix::fs::PermissionsExt as _;
use std::path::PathBuf;
use std::sync::Arc;
use std::sync::{Arc, Mutex};
use anyhow::{Context, Result};
use axum::{Json, extract::State, routing::get};
use hive_jobq_wire::GraphWire as _;
use serde::{Deserialize, Serialize};
use utoipa::{OpenApi, ToSchema};
use utoipa_axum::{router::OpenApiRouter, routes};
mod status;
/// Placeholder node payload for the swarm-level job graph — uninhabited on
/// purpose, and named `Swarm*` rather than the bare `NodeKind`/`Resource`
/// `hive-c0re::job_queue::model` already uses, so a grep for either doesn't
/// land on both crates. This wires the graph and its read-only endpoints;
/// giving it real variants waits on there being an actual job (agent
/// creation) to run. `WireNode` is trivially satisfiable on an empty enum
/// (`match *self {}`), so the wire machinery below is real and typechecked
/// today, with nothing yet to put in it.
#[derive(Clone, Debug)]
enum SwarmNodeKind {}
impl hive_jobq_wire::WireNode for SwarmNodeKind {
fn label(&self) -> String {
match *self {}
}
fn data(&self, _id: hive_jobq_wire::WireId) -> serde_json::Value {
match *self {}
}
}
/// Placeholder resource name — same rationale and same "no variants until a
/// real node needs one" shape as [`SwarmNodeKind`].
#[derive(Clone, Debug, PartialEq, Eq, Hash)]
enum SwarmResourceKind {}
impl hive_jobq_wire::WireResource for SwarmResourceKind {
fn name(&self) -> String {
match *self {}
}
}
/// Where the daemon binds, overridable via `SWARM_CONTROLLER_SOCKET`.
///
/// A compiled-in default is legitimate here and is *not* the mistake that
@ -68,6 +101,7 @@ fn socket_path() -> PathBuf {
(name = "health", description = "liveness probe"),
(name = "hives", description = "the swarm's hive directory"),
(name = "links", description = "swarm service quick links"),
(name = "jobq", description = "the swarm-level job graph"),
)
)]
struct ApiDoc;
@ -108,6 +142,11 @@ struct AppState {
/// that is merely *unreachable* still yields a reader, because
/// `async-nats` reconnects underneath it.
status: Option<Arc<status::StatusReader>>,
/// The swarm-level job graph. `std::sync::Mutex`, not `tokio`'s — every
/// lock scope below is synchronous (no `.await` while held). Always a
/// graph, never gated on the swarm queue: this is process state, not
/// something read over the network.
jobq: Arc<Mutex<hive_jobq::Graph<SwarmNodeKind, SwarmResourceKind>>>,
}
/// Env var the controller's NixOS module sets from
@ -251,6 +290,60 @@ async fn get_hives_status(
}
}
/// Query params for `GET /api/jobq/graph` — `?states=` narrows to root
/// groups in the named states, same shape `hive_jobq_wire::parse_states`
/// parses.
#[derive(Deserialize, utoipa::IntoParams)]
struct JobqGraphQuery {
states: Option<String>,
}
/// Every node of every root group in the swarm-level job graph. Nothing
/// filters "done and old" here — with zero nodes ever submitted there is
/// nothing to bound yet, so `graph.roots()` (everything) is the whole
/// roster passed to [`hive_jobq_wire::GraphWire::wire_snapshot`].
#[utoipa::path(
get,
path = "/api/jobq/graph",
params(JobqGraphQuery),
responses((status = 200, description = "every node of every root group, as generic \
`hive_jobq` graph nodes. `?states=` narrows to root groups in the named states.",
body = Vec<hive_jobq_wire::GraphNode>)),
tag = "jobq"
)]
async fn get_jobq_graph(
State(state): State<AppState>,
axum::extract::Query(q): axum::extract::Query<JobqGraphQuery>,
) -> Json<Vec<hive_jobq_wire::GraphNode>> {
let states = hive_jobq_wire::parse_states(q.states.as_deref());
let graph = state
.jobq
.lock()
.unwrap_or_else(std::sync::PoisonError::into_inner);
let roots: Vec<hive_jobq::NodeId> = graph.roots().map(|n| n.id).collect();
let nodes = graph.wire_snapshot(roots);
Json(hive_jobq_wire::filter_nodes_by_state(
nodes,
states.as_deref(),
))
}
/// Counts by lifecycle state over the same groups `/api/jobq/graph` serves.
#[utoipa::path(
get,
path = "/api/jobq/rollup",
responses((status = 200, description = "counts by lifecycle state", body = Vec<hive_jobq_wire::StateCount>)),
tag = "jobq"
)]
async fn get_jobq_rollup(State(state): State<AppState>) -> Json<Vec<hive_jobq_wire::StateCount>> {
let graph = state
.jobq
.lock()
.unwrap_or_else(std::sync::PoisonError::into_inner);
let roots: Vec<hive_jobq::NodeId> = graph.roots().map(|n| n.id).collect();
Json(hive_jobq_wire::state_rollup(&graph, roots))
}
#[tokio::main]
async fn main() -> Result<()> {
tracing_subscriber::fmt()
@ -334,6 +427,7 @@ async fn main() -> Result<()> {
hives: Arc::new(load_hives()),
links: Arc::new(load_links()),
status,
jobq: Arc::new(Mutex::new(hive_jobq::Graph::new())),
};
let (router, api) = OpenApiRouter::<AppState>::with_openapi(ApiDoc::openapi())
@ -341,6 +435,8 @@ async fn main() -> Result<()> {
.routes(routes!(get_hives))
.routes(routes!(get_hives_status))
.routes(routes!(get_links))
.routes(routes!(get_jobq_graph))
.routes(routes!(get_jobq_rollup))
.split_for_parts();
// Just the JSON, not the UI — Swagger UI itself is nginx-hosted from
// the nix store (see the module doc comment above). `api` is