Closes 2718.
The operator has been going through agent dirs by hand with ncdu,
deleting 20+GB target dirs. Agents had no way to know they were the
ones sitting on the space.
New `disk_watch` module in the harness: every 15 minutes it statvfs's
the filesystem backing the agent's state dir and, past 80%, raises a
keyed `disk` todo telling the agent to free space — with the operator's
rules inline: only delete things that are actually big, build output
first, and never delete something still needed, ask for more space
instead.
Over threshold it also walks the agent's own tree (`/agents/<label>`
plus `$HOME`) and names the directories worth looking at, so the todo
says where the bytes actually went rather than just that the disk is
full. The walk is bounded on every axis — entry budget, recursion cap,
report depth — pinned to the state dir's device so it can't wander into
`/nix` or the shared bind mounts, and it does not traverse symlinks. It
reports the deepest oversized directory on each branch, so the agent
gets pointed at `<workspace>/target` rather than at `/agents/<label>`.
Anti-nag is the whole design constraint. The todo is keyed, and the
summary is deliberately stable: the percentage is bucketed to 5 points
and no raw byte counts appear anywhere in it. An unchanged situation
re-upserts as `changed == false` and never fires the wake, so a disk
that has been steady at 89% for a week sits quietly in the loose-ends
list; crossing into a new bucket speaks up once. Dropping back under
the threshold clears the row.
Harness-local by construction, per the operator's call that this gets
no core wiring: hive-c0re cannot push a todo at all (the store and its
wake live inside the container), and running in-process means this
skips even the in-agent socket and calls `Todos::upsert` directly.
Worth recording, since it shaped the scope: btrfs does NOT fold qgroup
limits into statfs. Measured with quota counting enabled and a 20G
limit set on a real subvolume, statvfs returns byte-identical whole-FS
numbers for that subvolume, an ordinary agent dir, and the root. So
this watches host-FS pressure, which is valid before and after the
planned subvolume migration; per-agent quota awareness would need the
limit handed to the agent explicitly.
Node gains started_at/finished_at (chrono DateTime<Utc>, serialized
RFC 3339 on the wire per hive_sh4re::wire_time) plus error (String).
Graph::set_state self-stamps started_at on the first Running transition
and finished_at on the first terminal one, via an internal now_utc()
clock (keeps settle/complete signatures stable). Outcome::Failed(String)
carries the failure reason, set on the terminal transition.
hive-c0re complete_node builds Outcome::Failed(msg); its node_rt
side-table stays i64 for now (double-write) until #2637 reads the Node.
Toward #2637: the jobq graph becomes the source of truth for per-node
lifecycle so the queue can be sent to the client as-is.
Replace the in-tree scheduler with the domain-agnostic hive-jobq crate
(merged in #2615): parent-axis grouping + borrow/subtree-reservation
resource model + roll-up completion (State::Finishing).
Host adaptation:
- NodeSpec gains an explicit `parent` axis; templates declare grouping +
sibling ordering directly (deps order execution, parent groups a subtree
whose resource the descendants borrow).
- Rebuild is a nested two-root subtree: Prebuild (root, owns the build slot
for the whole subtree, lease-exempt) -> StopForUpdate (child, owns the
agent lease) -> Swap/PostSwap (children, borrow both); Reconcile is a
separate top-level root (AfterAny Prebuild) so it survives the cancel-
cascade of any failed step (recovery-start invariant) and converges to
the persisted `wanted` on a fresh lease. This is the multi-root
correction to the single-root-chain sketch: node0=root broke lease-
exemption (hoisting the lease onto Prebuild) and recovery-reconcile
(root failure cancels all children).
- Spawn / perm-change / power-ops (stop/start/restart) group-rooted the
same way; per-agent power-op subgraphs stay independent roots so a
multi-agent DAG runs them concurrently, each on its own lease.
- insert_group honours the explicit parent axis (no lease hoisting); the
DAG terminal node deps AfterAny on every group root and runs once the
whole op rolls up. Drop the old Graph::add_dep terminal wiring.
36/36 job_queue tests, full hive-c0re suite green, clippy --all-targets.
New crate hive-screen-mcp: a stdio MCP bridge activated automatically
when an agent has hyperhive.gui.enable = true. Provides five tools:
- screenshot — grim → saves PNG, returns path for Read tool
- type_text — wtype → Unicode text input (no daemon)
- key_press — ydotool key → combos like ctrl+c, super+l
- mouse_move — ydotool mousemove --absolute
- mouse_click — ydotool click, optionally with prior move
New nix/agent-modules/screen.nix: wires the MCP bridge into
extraMcpServers.screen; adds grim + wtype to systemPackages. Adds
hyperhive.gui.screenInput option (default false) which enables the
ydotoold daemon + ydotool for mouse/keyboard injection via /dev/uinput.
screenshot and type_text work without screenInput. key_press,
mouse_move, and mouse_click return a ydotool error until ydotoold is
running and /dev/uinput is accessible in the container.
Per mara's direction — validate ids as they enter the graph so internal
iteration can trust every id the graph holds; the generational route for
removal comes later. Adds GraphError; insert() now rejects a dangling
Dep::Node / parent id (it is fallible); validate() checks all internal id
references resolve and that next_id is past the largest existing id;
deserialization runs validate() via #[serde(try_from = "GraphData<N>")], so
a loaded graph can never carry a dangling reference. 5 new tests; serde_json
added as a dev-dependency for the round-trip cases.
First step of extracting the job-DAG queue into a domain-agnostic
`hive-jobq` library, per the operator's v2 design: one persistent
graph, named-counter resources, recursive node groups, opaque stable
node ids, guard-object locks, a slot-filling scheduler.
This commit lands only the data model, so the shape can be reviewed
before the machinery is built on it:
- NodeId: opaque, stable, monotonic; group membership is a parent
edge, not encoded in the id (the 1/1/2 hierarchy is a derived UI
label).
- ResourceName, Dep (Node | Resource{name,count}), State.
- Node<N>: caller-defined payload N so the library stays
container-agnostic.
- Graph<N>: insert (mints stable ids), node lookup, children,
recursive group-terminal check. Retains completed groups (no
pruning in v1).
The resource-acquisition machinery (atomic all-or-nothing acquire),
the recursive-lock guards, and the scheduler loop are follow-ups.
Tests cover id minting, group terminality, and state terminality;
clippy + rustdoc clean.
Now that matrix-sdk 0.18 is on main, reqwest 0.13.1 is already in the
tree transitively. Point the workspace crates at it directly.
reqwest 0.13 renamed the rustls feature set:
- rustls-tls -> rustls
- rustls-tls-native-roots -> rustls-native-certs
- (webpki-roots is now a separate feature)
hive-forge keeps its dual-trust story (system/native store for the
hive CA + bundled Mozilla roots for public CAs) by enabling
rustls-native-certs + webpki-roots explicitly.
forgejo-api 0.11 resolves cleanly against reqwest 0.13 (no conflict).
rusqlite 0.40 is intentionally NOT bumped here: matrix-sdk-sqlite 0.18
still pins rusqlite 0.37, so 0.40's libsqlite3-sys 0.38 would hit the
links="sqlite3" single-owner conflict. Deferred until upstream moves.
Bumps the feature (major) versions that update cleanly without breaking the
build: indicatif 0.17->0.18, tower-http 0.6->0.7, hmac 0.12->0.13,
sha2 0.10->0.11. Only adaptation needed: import hmac's KeyInit trait in
webhook_secret (new_from_slice moved from Mac to KeyInit in hmac 0.13).
Held back (require dedicated code-change PRs, out of scope for a
non-breaking bump):
- reqwest 0.13: renames the rustls-tls feature and conflicts with
forgejo-api 0.11 + matrix-sdk 0.14 which pin reqwest 0.12.
- rusqlite 0.40: libsqlite3-sys 0.38 clashes with matrix-sdk-sqlite 0.14's
0.35 (single links=sqlite3) — coupled to the matrix-sdk bump.
- rmcp 2.2, matrix-sdk 0.18: major API rewrites across the MCP/matrix crates.
Refreshes ~90 semver-compatible dependency versions (patch/minor within
the existing Cargo.toml constraints); Cargo.lock only, no source changes.
Major bumps held back by our constraints (rmcp 2.x, matrix-sdk 0.18,
reqwest 0.13, rusqlite 0.40, tower-http 0.7, sha2 0.11) are left for
separate, code-touching follow-ups.
hive-c0re already samples each agent container's cgroup load for the
dashboard (stats/container_stats.rs); this rides those gauges out to the
configured OTLP endpoint, reusing the existing services.hyperhive.otel
config (endpoint + auth header) — no new toggle.
- New stats/otel_metrics.rs: exports via the OpenTelemetry Rust SDK (same
crates as hive-metric) with the semconv container.* metric names +
container.name attribute so off-the-shelf OTel/Grafana dashboards work,
plus the hive agent label. container.cpu.time (counter, s, from cumulative
cpu.stat usage_usec), container.memory.usage, container.memory.usage.limit;
memory peak / on-disk storage / instantaneous cpu percent stay hyperhive.*
custom (no semconv equivalent). Observable instruments read a shared
snapshot an async task refreshes (gather() is async; SDK callbacks sync).
- container_stats: expose cpu_time_usec (cumulative) on ContainerResource.
- The OTLP auth header is loaded onto hive-c0re's own unit via systemd
LoadCredential and read from $CREDENTIALS_DIRECTORY/otel-headers.
- docs/observability.md documents the host-emitted semconv metrics.
Host-side export, so it covers containers even when their agent is idle.
Split the priv-socket wire types (PrivRequest/PrivResponse/PrivEvent and
friends) out of hive-sh4re into their own hive-priv-sock crate, mirroring
the existing hive-host-sock split. hive-priv — the root-privileged
helper — now depends on just this narrow protocol crate instead of the
much larger daemon-shared crate, shrinking its dependency surface and
making the privsep boundary easier to audit. No server/client
implementation lives here, only the wire contract; hive-c0re still
depends on hive-sh4re directly for everything else.
Both webhook registrations (knowledge push + config-PR pull_request) now
use the public hive domain instead of loopback:
https://<HYPERHIVE_HIVE_DOMAIN>/webhook/{knowledge,config-pr}
This routes deliveries through the gateway, bypassing the Forgejo SSRF
guard that blocked loopback delivery and silently broke the config-PR
merge flow since launch.
Changes:
- webhook_secret: new module — auto-generate + persist a 32-byte HMAC
secret to STATE_ROOT/webhook-secret on first startup; verify
X-Hub-Signature-256 on every incoming webhook POST (HMAC-SHA256).
- forge/mod.rs: ensure_config_pr_webhook now takes hive_domain +
webhook_secret; sets secret in Forgejo hook config.
- workers/knowledge.rs: ensure_webhook same update.
- dashboard/webhook.rs: both handlers read raw Bytes first, verify HMAC,
then parse JSON. Returns 401 on signature mismatch.
- dashboard/mod.rs: AppState carries webhook_secret; serve() takes it.
- main.rs: load/generate secret at startup; pass to registration tasks
+ dashboard; add 5-minute config-PR polling fallback task.
- forge/config_pr_poll.rs: new — scan agent-configs/* for open PRs with
no pending MergeConfigPr approval; queue them. Idempotent.
- stores/approvals.rs: has_pending_merge_config_pr() for poll dedup.
- nix/modules/hive-gateway.nix: remove dashboardAuth from /webhook/
location (HMAC replaces basic auth for webhook endpoints; Forgejo
cannot send HTTP Basic credentials with webhook deliveries).
reqwest has no UDS transport, so unix: upstreams dial the socket
directly with a raw hyper/1.1 client per request instead. http(s)://
upstreams are unaffected (still go through the existing reqwest path).
Adds hyper (client, http1), hyper-util (tokio IO adapter), and
http-body-util as direct hive-ag3nt dependencies - all three were
already present transitively via reqwest, this just uses them
directly for the new code path.
jobs are now DAGs of primitive nodes (prebuild, stop-for-update, swap,
reconcile, signal, drain, ...) driven by one scheduler with N build
slots + per-agent lifecycle leases. per-agent power intent (wanted
up/offline) is durable in agent_power.sqlite; Reconcile nodes converge
observed state to it. kills the graceful-stop watcher thread, the
deferred-start follow-up, and the cascade pre-enqueue (fan-out on
MetaLock completion instead). tracker: #2166
