The clippy check's comment described `-D warnings -A clippy::pedantic` — the
`-A` half dropping pedantic from the CI gate — but the args were only
`-D warnings`, so pedantic was hard-denied contrary to the doc. Operator
call: pedantic should be gated. Encode that as the single source of truth:
set the workspace lint `pedantic = deny` (errors locally and in CI), and
rewrite the checks.nix comment to match. Args unchanged; `-D warnings` still
gates rustc + non-pedantic clippy warnings. No new failures — the tree was
already pedantic-clean under CI's `-D warnings`, which denied pedantic.
hive-core now fetches each agent's config as a forge-hosted flake input
(git+http://<forge>/agent-configs/<n>.git). Add a git credential helper
(git-credential-hive-forge) that reads the live forge-core admin token on
every invocation and authenticates as the forge core user, wired via the
[credential] stanza in hive-core's $HOME/.gitconfig + on the service PATH.
Reading the token file live means zero stale copies and no resync on
rotation; no token ever lands in a flake URL or lock. Mirrors the existing
github.nix credential-helper pattern.
package-lock.json was updated in 9edd3750 (dompurify/marked/esbuild
bumps) but npmDepsHash in nix/packages/frontend.nix wasn't recomputed,
breaking the frontend FOD build on main (hash mismatch, fixes#2539).
Recomputed via prefetch-npm-deps frontend/package-lock.json.
Completes #2415. Registration no longer gates container@hive-ci start:
- Retire the host-side hive-ci-prefetch.service (+ its 100-line script and
the now-dead coreTokenPath/containerRoot let-bindings) — the forge round-trip
it did on the boot path now lives in hive-c0re (forge/ci_runner.rs), run out
of band during the startup sweep.
- Drop the container@hive-ci TimeoutStartSec = mkForce 180s band-aid that
papered over that boot-path wait tripping the nspawn start timeout (the ~60s
restart loop of #2410).
- gitea-runner-hive gains an ExecStartPre precond (ahead of the nix-daemon
wait) that fails fast unless it is already registered (.runner present) or a
real, non-placeholder token is in place — so missing creds just hold the
runner down instead of blocking the container. Restart=on-failure (no start
limit) self-heals it once hive-c0re writes the token; c0re's explicit restart
is the primary path.
- Set HYPERHIVE_FORGE_CI_ENABLED=1 on hive-c0re.service so the sweep registers.
The tmpfiles TOKEN=placeholder seed + read-only bind-mount are unchanged; the
core admin token still never enters the container.
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.
hive-ci and hive-forge both bind the runtime-generated hive CA cert
read-only and order their container@ unit after hive-tls-ca.service so
the bind source exists before nspawn sets the mount up — the same
bind-mount + ordering + rationale duplicated verbatim in two modules.
Extract that language-agnostic half into a pure helper,
nix/host-modules/lib/hive-ca-trust.nix, taking a container name and
returning { useSelfSigned, caContainerPath, bindMount, containerOrdering }.
The per-runtime consumption stays at each call site (hive-ci's additive
NODE_EXTRA_CA_CERTS, hive-forge's Go SSL_CERT_FILE concat). hive-ci folds
containerOrdering into its existing mkMerge alongside the TimeoutStartSec
bump.
The helper is a pure function, not a module: host-modules/default.nix is
an explicit aggregator (not a glob) and the docs eval imports that same
aggregator, so the lib/ file is never picked up as a module. A third
outbound-TLS-trusting container no longer means a third copy-paste.
Per mara's feedback on PR #2407 ("better: you can also provide url in
dashboard, same as with matrix, no host config"), drops
services.hyperhive.extraForges and the admin-API mint/revoke flow
entirely. The operator now creates a token on the external forge
themselves and pastes a label + base URL + access token into the
dashboard's FORGES tab, the same shape as the GitHub PAT flow plus the
base-URL field from the matrix extra-account flow. hive-c0re only ever
writes/deletes two local files per account (forge-<label>-token,
forge-<label>.json sidecar for the URL) via hive-priv — no remote
account creation, no admin token, no revoke-on-the-remote-side, no nix
config to enumerate.
- nix/host-modules/hive-forge/default.nix: removed the extraForges
option, its label-format assertion, and the HYPERHIVE_EXTRA_FORGES
env forwarding.
- hive-c0re/src/forge/extra.rs: deleted (REST admin-API provisioning,
no longer needed).
- hive-c0re/src/dashboard/extra_forges.rs: GET /api/extra-forges?
agent= lists an agent's stored forges by scanning its state dir
(mirrors matrix_accounts.rs's filename-scan listing), POST
/api/extra-forge-account (agent/label/base_url/token/
action=add|remove) stores or removes an account.
- hive-sh4re/priv_proto.rs + hive-priv/main.rs: new
WriteAgentExtraForgeAccount/DeleteAgentExtraForgeAccount priv
requests (adds base_url, writes/deletes a JSON sidecar alongside the
token).
- hive-c0re/src/priv_client.rs: matching wrapper functions.
- frontend/packages/dashboard/src/credentials.{html,js}: FORGES tab is
a per-agent list + add-account paste form (label/base_url/token), no
grant/revoke-from-catalog UI.
- docs/web-ui/dashboard.md: FORGES tab section rewritten.
Supersedes the design in PR #2407 (already approved+green on the old
admin-API model) — opening as a fresh PR against the same issues
rather than force-pushing over the approved one.
gitea-runner registration (hive-ci-prefetch, host-side) sits on the
boot-critical path -- nspawn readiness is gated on the runner
registering, itself a forge + core-token round trip that can wait up
to 60s for the core token. The default ~60s TimeoutStartSec can trip
mid-register (especially right after a .runner purge, since every
boot re-registers from scratch), killing the half-started container
and triggering a restart loop until the token/forge settle. Bump to
180s so one register attempt has room to finish.
mara: "no, do split packages for the separate daemons" — instead of
grouping all daemon/harness/MCP bins into one hyperhive-daemon
derivation, each bin (hive-c0re, hive-priv, hive-agent,
hive-agent-mcp, hive-agent-wake, hive-bash-daemon, hive-bash-mcp,
hive-matrix-daemon, hive-matrix-mcp, hive-metric) is now its own
named package + overlay output, matching the existing hivectl /
hive-forge split. harness-base.nix's ExecStart/command lines now
point at the specific pkgs.hive-* each ExecStart needs, and
environment.systemPackages only carries the two bins actually
PATH-looked-up in-container (hive-agent-wake, hive-metric) instead
of the whole bundle.
Agent containers pulled in `pkgs.hyperhive` (packages.default: daemonBins +
hivectlPkg + hiveForgePkg) via environment.systemPackages and every
harness ExecStart. That shipped hivectl (a host-admin CLI dialing the
host admin socket, unreachable from inside a container, wrapped with
wireguard-tools for `hivectl wg`) plus a redundant second hive-forge
copy (already provided separately via hive-forge-tools.nix) into every
single agent's closure.
Expose the existing daemonBins split as packages.<system>.daemonBins
(mara: 'there should be split nix outputs already, use those' — no new
grouping derivation) + a hyperhive-daemon overlay entry, and repoint
harness-base.nix's systemPackages entry and all agent-run ExecStarts
(hive-agent, hive-agent-mcp, hive-bash-daemon, hive-matrix-daemon) plus
the bash/matrix MCP bridge commands at it instead of the full bundle.
nix flake check: daemonBins + agent-base-toplevel + ruth-toplevel all
evaluate cleanly (docs derivation failure is a preexisting dirty-tree
issue, unrelated).
- docs/network.md: LOCAL_ADDRESS is now empty (not deterministic-IP);
containers use dhcpcd + dnsmasq DHCP pool, not hash-derived static IPs
- nix/modules/hive-ci.nix: 'service containers' → 'all containers'
in the dnsmasq DHCP pool comment (agents also use the same pool)
All agent containers now receive their bridge IP dynamically via DHCP
from the dnsmasq pool instead of a hash-derived static address:
- nix/templates/harness-base.nix: networking.useDHCP = true
- nix/modules/hive-gateway.nix: expand DHCP pool to full usable range
(.2 to .254 on /24) — was last-14-IPs-only
- hive-sh4re/src/priv_proto.rs: remove agent_ip from NetworkIsolation
- hive-c0re/src/lifecycle/mod.rs: drop agent_network_ip + DHCP_POOL_SIZE
- hive-c0re/src/lifecycle/host_config.rs: remove agent_network_ip call
- hive-priv/src/main.rs: LOCAL_ADDRESS= empty (DHCP assigns IP);
HOST_ADDRESS still set so nixos-container installs default route
before the DHCP lease arrives
- nix/dhcp-pool-size: deleted (no longer needed)
The nix/dhcp-pool-size single-source-of-truth file and all associated
Rust/Nix dual-constant plumbing are gone — there is no static map.
bridge_gateway_ip() is retained (still needed for HOST_ADDRESS).
Closes#2363
The DHCP pool exclusion is now by construction (agent_slots = usable -
dhcpPoolSize), not a secondary-hash remap. Update two comment blocks
that still referenced the old approach.
Move the DHCP pool size constant out of the two separate definitions
(Nix literal + Rust const) into a shared data file: nix/dhcp-pool-size.
- nix/dhcp-pool-size: new file, contains '14'
- hive-gateway.nix: reads via builtins.readFile + toIntBase10
- lifecycle/mod.rs: parses via include_bytes! const block at compile time
Cargo automatically tracks include_bytes! as a file dependency so a
change to nix/dhcp-pool-size triggers recompilation without build.rs.
Add a DHCP pool to the gateway's dnsmasq so bridge-attached service
containers (hive-ci and future equivalents) get their addresses from
a proper DHCP server instead of a brittle static derivation.
gateway (hive-gateway.nix):
- Add IPv4 arithmetic helpers (ipToInt, intToIp, pow2) to compute the
DHCP pool range at nix eval time from bridgeIp + bridgePrefixLength.
- Reserve the last dhcpPoolSize (14) usable host addresses as the DHCP
pool (e.g. .241-.254 on a /24 with 10.42.0.0 network).
- Add dhcp-range and dhcp-leasefile to the dnsmasq settings block.
The pool is active whenever services.hyperhive.network.enable is true.
hive-ci (hive-ci.nix):
- Remove the ciBridgeIp / ciBridgeOctets static derivation and the
brittle top-of-/24 comment block.
- Switch networking.interfaces.eth0 to useDHCP = true so hive-ci gets
its address from the gateway DHCP pool.
lifecycle (mod.rs, tests.rs):
- Add DHCP_POOL_SIZE = 14 constant (must stay in sync with
dhcpPoolSize in hive-gateway.nix).
- Remap agents whose FNV-1a hash falls in the DHCP pool into the
agent-only window [2, dhcp_start - 1]. Only the rare agent whose
name hashes into the pool is affected; all others keep their IPs.
- Update and extend tests: agent range is now .2-.240 on /24;
add agent_network_ip_never_in_dhcp_pool covering 18 agent names.
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).
