Watch
0
0
Fork
You've already forked hyperhive
0

swarm-builder: build each agent config PR head and post a commit status

A new `swarm-builder` service, in its own container on the host that sets
`deploy.builder.enable` (defaulted on by `allSwarmServices` where the host
has the builder's store leaf, like the review classifier).

When an `agent-configs/<agent>` PR is opened, reopened or synchronized,
swarm-controller queues a `build_agent` jobq node. The node sets a
`swarm-builder` commit status to pending on the PR head, sends
`{agent, rev, pr}` to the builder over NATS request-reply
(`$SWARM.builder.build`), and writes success (store path), failure (last
line nix printed) or error (no builder / no reply) on the head. The node
fails when the build does, so the Jobs page matches the PR.

The builder builds nix/builder/check-flake's `check` with every input
overridden: hyperhive and nixpkgs to the host's own
`c0re.hyperhiveFlake`/`c0re.nixpkgsFlake` (what a hive there renders into
its meta flake), and the agent to the PR head on the forge. Evaluation is
pure and runs in the container; builds go to the host nix daemon as an
untrusted client. The private config repo is fetched with the token of a
new read-only `builder` forge account (in the `hives` team), which the
controller mints into `swarm/services/builder/forge/token`; git gets it
from a credential helper reading the child's environment, so it is in no
argument or URL.

Credentials: a `builder` store leaf minted by glue-bao-tls on the store
host, a `swarm-builder` reader policy (forge token + its own OIDC client
secret), an authelia machine client, and a `--builder-client` grant in the
queue's auth-callout responder. The controller gains a grant on the token
path and the secret publisher a deny on it.

This is the agent-config build check atlas scoped on #5081 (slice 3):
nothing is pushed to a cache, nothing is written to KV, and a red status
gates no deploy. The check composes agent-base + the config without the
values a hive adds (CA, docs, claude-code path, ports, input follows), so
it can disagree with a hive's own build.
This commit is contained in:
atlas 2026-10-12 22:04:51 +02:00
commit b68b74d687
43 changed files with 2452 additions and 42 deletions

View file

@ -0,0 +1,75 @@
//! The swarm builder's builds: the subject swarm-controller sends a build on,
//! and what travels each way.
//!
//! Three crates name these: the controller sends the request, the builder
//! answers it, and the auth-callout responder grants the one subject to each.
use std::time::Duration;
use serde::{Deserialize, Serialize};
/// The subject every build goes to. The controller is its one publisher and
/// the builder its one subscriber; the builder answers on the request's reply
/// subject.
pub const SUBJECT: &str = "$SWARM.builder.build";
/// The queue group the builder subscribes in, so a second instance shares the
/// requests instead of running each one twice.
pub const QUEUE_GROUP: &str = "builder";
/// The longest one build may take before the builder stops it and answers
/// with a timeout.
pub const BUILD_DEADLINE: Duration = Duration::from_hours(1);
/// How long the controller waits for the reply, and the queue lets the
/// builder answer: the build at its deadline, plus a minute for the rest.
pub const REPLY_DEADLINE: Duration = Duration::from_mins(61);
/// One build: an agent's config repo at one commit, on the builder's own
/// agent base.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct BuildAgent {
/// The agent, which is also its config repo's name.
pub agent: String,
/// The commit to build.
pub rev: String,
/// The PR the commit is the head of; its `refs/pull/<pr>/head` is the ref
/// the builder fetches the commit through.
pub pr: u64,
}
/// The builder's answer to one [`BuildAgent`].
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(tag = "state", rename_all = "kebab-case")]
pub enum Reply {
/// The config built; `out_path` is the system closure's store path.
Built { out_path: String },
/// The build did not finish: the last lines `nix` printed, or why the
/// builder could not start it.
Failed { error_tail: String },
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn the_wire_tags_the_reply_by_state() {
let json = serde_json::to_value(Reply::Built {
out_path: "/nix/store/x".to_owned(),
})
.expect("encodes");
assert_eq!(json["state"], "built");
assert_eq!(json["out_path"], "/nix/store/x");
let json = serde_json::to_value(Reply::Failed {
error_tail: "e".to_owned(),
})
.expect("encodes");
assert_eq!(json["state"], "failed");
}
#[test]
fn the_reply_deadline_outlasts_the_build() {
assert!(REPLY_DEADLINE > BUILD_DEADLINE);
}
}