mara/hyperhive
mara/hyperhive
Watch
0
0
Fork
You've already forked hyperhive
 0
Code Activity
hyperhive/docs/4-implementation/networking/network.md

57 KiB
Raw Blame History

hive-network

The bridge every container on a host attaches to, and the private-netns isolation of agent and CI containers behind it. It comes up on any host that runs agents or swarm services: services.hyperhive.network.enable defaults to off, and the modules that need the bridge set it with mkDefault true — the hive controller, CI, the hive collector, and the gateway's resolver, which every swarm service turns on. Configured via services.hyperhive.network.*.

Agent containers always run in a private netns.

Network map

One picture of the whole hive — two planes: infra containers share the host netns and bind host ports directly; compute containers (agents + CI) each get a private netns behind the bridge. The unix-socket control plane rides the VFS, and none of that touches it.

                        internet
                           │  uplink NIC — NAT MASQUERADE for the
                           │  bridge subnet (10.42.0.0/24 default)
┌──────────────────────────┴─────────────────────────────────────────┐
│ host netns — the host itself plus gateway / forge                  │
│                                                                    │
│   nginx :80/:443                                  [hive-gateway]   │
│   dnsmasq 10.42.0.1:53 (DNS) + :67 (DHCP)         [hive-gateway]   │
│   squid 10.42.0.1:3128 (egress proxy)             [hive-squid]     │
│   forgejo :3000 http, :2222 git-ssh               [hive-forge]     │
│   hive-c0re dashboard 127.0.0.1:7000              (host service)   │
│   wg-hive :51820/udp — swarm mesh, when enabled   (host iface)     │
│                                                                    │
│                  hive-br0   10.42.0.1/24                           │
│              ┌──────────┼──────────────┐                           │
└──────────────┼──────────┼──────────────┼───────────────────────────┘
          vb-h-<a>    vb-h-<b>      vb-<runner>     veth pairs
               │          │              │
          ┌────┴────┐ ┌───┴─────┐ ┌──────┴──┐   one private netns
          │ agent a │ │ agent b │ │ runner  │   each; eth0 leases
          │  eth0   │ │  eth0   │ │  eth0   │   from the DHCP pool
          └─────────┘ └─────────┘ └─────────┘
container netns IPv4 listens / reached via
gateway (host) host host addresses nginx :80/:443 (every vhost); dnsmasq bridgeIp:53 + DHCP :67 on the bridge
egress proxy (host) host bridgeIp squid bridgeIp:3128; agents with a credential send HTTP and HTTPS through it
hive-forge host (shared) host addresses forgejo :3000 http, :2222 git-ssh; fronted by the forge.<swarm-domain> vhost
hive-matrix private, veth on bridge DHCP pool tuwunel on UDS /run/hive-matrix/tuwunel.sock → the matrix vhost; out to store, SSO, peers
swarm-ci-runner private, veth on bridge DHCP pool outbound only (runner → forge); no inbound surface
swarm-review-classifier private, veth on bridge DHCP pool outbound only (classifier → queue, secret store, swarm collector); no inbound surface
h-<agent> private, veth on bridge DHCP pool web UI via UDS /run/hive-agent/<name> → nginx sub-path; in-container UI port hashed 8100–8999

The flows, end to end:

  • DHCP — agent dhcpcd broadcasts on eth0 → veth → bridge → host firewall (udp 67 hole) → dnsmasq pool → lease + router option. dhcpcd runs with noipv4ll and waits for an IPv4 address, so network-online.target means a lease, never a 169.254.x fallback; units that look up a swarm name at start order after it.
  • DNS — agents and the service containers query bridgeIp:53; dnsmasq answers hive zones authoritatively with the bridge IP, everything else forwards to the host's resolvers (see Resolver behaviour below). Each container points its own resolv.conf there, and one that instead inherits the host's resolves no swarm name at all — those records exist only on the bridge.
  • HTTP — forge. and chat. (under swarm.domain) plus the hive's own dashboard name resolve to the bridge IP, land on nginx :80/:443, and proxy to forgejo :3000, tuwunel's unix socket, hive-c0re 127.0.0.1:7000, or a per-agent UI unix socket.
  • Internet egress — agent default route points at the bridge IP; the host forwards + masquerades out its uplink.
  • Swarm — peer hives connect over the wg-hive WireGuard mesh and reach each other's gateway/forge across it (docs/2-handbook/swarm/).
  • Control plane (no network) — per-agent broker socket /run/hive/mcp.sock, privileged helper /run/hive/priv.sock, operator admin /run/hyperhive/host.sock, and the per-agent UI sockets under /run/hive-agent/ are unix domain sockets bind-mounted through the VFS; private netns doesn't affect them.

Container shape (where dnsmasq lives)

dnsmasq runs on the host itself, next to nginx — both come from the gateway module, one front door for DNS and HTTP. Its bind-interfaces listener sits on bridgeIp, on the host's bridge interface.

Configuration

{
  services.hyperhive = {
    swarm = {
      domain = "myswarm.example";
      hives.firsthive = { }; # -> domain = firsthive.myswarm.example
    };
    deploy.hive-controller = {
      enable = true;
      hiveName = "firsthive";
    };
    # network.bridgeIp = "10.42.0.1"; # default
  };
}

The hive domain (services.hyperhive.domain, which the resolver answers for) comes from this hive's entry in the swarm directory: swarm.hives.<hiveName>.domain, default <hiveName>.<swarm.domain>. Eval fails until swarm.domain and that entry exist (2-handbook/swarm/README.md § Hive identity config).

Bridge addressing

Default subnet is 10.42.0.0/24, host-side gateway at 10.42.0.1. RFC 1918 space, unlikely to clash with operator's existing setup; override bridgeIp + bridgePrefixLength if a different range is already in use. /24 gives 254 usable per-agent addresses — enough for any single-host hive; bigger swarms or tighter addressing schemes pick their own.

Resolver behaviour

dnsmasq is authoritative for the hive's own zone (<hive-domain>) plus whatever swarm-service names this host contributes via gateway.localNames — forge.<swarm-domain> and chat.<swarm-domain> (matrix) when this host runs those services, auth.<swarm-domain> when it runs authelia, and hive-collector.hive on every hive — answering each with the bridge IP (where nginx is reachable, and where the hive's own collector listens). Note forge and matrix are swarm-domain names, not sub-domains of <hive-domain>: a swarm runs one forge and one homeserver, so their names belong to the swarm rather than to whichever hive happens to host them.

Every domain answered this way is also local= for dnsmasq, so the other record types (AAAA, TXT, MX, HTTPS) for those names get an empty answer from dnsmasq itself instead of waiting on the host's upstream resolver. Without it, dnsmasq 2.86 and later send them upstream, and a host whose upstream doesn't answer leaves each lookup to its client's own timeout.

dnsmasq forwards everything else to the host's own resolvers: it runs on the host and reads the host's /etc/resolv.conf directly. Containers don't need to know the upstream — they query the bridge IP and dnsmasq does the right thing per-name.

No fallback server= exists, by design: dnsmasq queries all known upstreams in parallel, so a hardcoded public resolver would take a share of normal traffic, not just cover the gap.

dnsmasq runs on the host and reads the host's /etc/resolv.conf directly, so a network change (new router, new lease, laptop moving networks) reaches it the moment openresolv rewrites the file. Nothing needs synchronising, and no unit watches for it.

bind-interfaces + interface = [ bridgeName "lo" ] means the listener only accepts queries from the bridge interface (plus lo for container health-checks). External hosts can't reach it — no DNS-amplification surface even when the operator opens port 80 for gateway HTTP.

resolveLocalQueries = false keeps dnsmasq out of the host's own resolution stack — the host's resolver (systemd-resolved, plain glibc nss, dnscrypt-proxy, etc.) keeps doing whatever the operator configured. The hive resolver is purely for inbound queries from agent containers.

Firewall posture

networking.firewall.interfaces.<bridge>.allowedUDPPorts = [ 53 67 ] networking.firewall.interfaces.<bridge>.allowedTCPPorts = [ 53 80 443 ]

  • Port 53 opens the resolver on the bridge interface only. Other interfaces stay closed. The hive resolver isn't an external-facing service.
  • Port 67 (UDP) admits DHCP requests to the dnsmasq pool. dnsmasq receives DHCP via a regular UDP socket (it doesn't use a netfilter-bypassing raw socket), so the hole is mandatory — without it containers never get a lease and fall back to 169.254.x.x.
  • Ports 80 and 443 let isolated agents reach nginx (on the host) for the forge sub-domain, per-agent UI proxies, and any other HTTP services.

The host firewall is the only firewall. The swarm service containers that share the host netns (forge, bao, the metrics and log stores) run with networking.firewall.enable = false (nix/container-modules/swarm-container.nix): a NixOS firewall inside a shared-netns container runs against the host ruleset — at container boot its firewall-start flushes the nixos-fw chains, rebuilds them from the container's (empty) port list, and deletes the host's nixos-nat-* chains without recreating them, silently wiping the bridge holes above plus the agents' NAT. Private-netns containers (every container with privateNetwork = true: agents and the service containers on the bridge) keep their own firewall — it's scoped to their namespace.

Reaching host services (exposeHostPorts)

By default agents can only reach the host on 80/443 (+53 DNS), so a host-side service on another port — for example a dev OTLP collector you want agents to reach directly — is unreachable. (hyperhive's own services need none of this: a module that means to be reachable from agents opens its own port here, which the hive collector and nats.enable both do. Note that deploy.swarm-otel.upstream.endpoint is the upstream, which no agent ever dials. See docs/4-implementation/scheduler/observability.md.)

services.hyperhive.network.exposeHostPorts = [ 4318 ]; opens each listed TCP port P on the bridge-interface allowedTCPPorts, so an agent can connect to <bridgeIp>:P (for a dev OTLP collector, point its endpoint at http://<bridgeIp>:4318, default http://10.42.0.1:4318). The hive's own collector needs none of this: agents dial it by otel.collector.host.

This is firewall-only: the host service must bind an address reachable from the bridge — 0.0.0.0 or the bridge IP — not loopback only. The bridge→127.0.0.0/8 DROP rule (below) still applies, so a service bound to 127.0.0.1 only stays unreachable; rebind it to 0.0.0.0.

The port is reachable by every agent on the bridge subnet (like DNS/gateway), so only expose services safe for any agent to reach.

Egress proxy (squid)

Every hive host runs squid in the host netns, listening on <bridgeIp>:3128 (nix/host-modules/hive-squid.nix). The port is open on the bridge interface only. An agent sends its outbound HTTP and HTTPS through it once its hive says the proxy is ready and it has fetched its proxy credential (see Agents and the proxy below). A program that ignores the proxy variables still reaches squid when it speaks HTTP on port 80 or TLS on port 443 and its agent is restricted or in log mode: the host redirects that traffic into squid's intercept ports (see Intercepted web traffic below). Everything else goes straight out through the NAT, as before: the unrestricted agents, the service containers and every other port and protocol. By default nothing blocks that. The host logs the direct connections of the restricted and log agents, and egress.forward = "reject" refuses them (see Agent bridge egress log below).

Squid filters on the destination hostname: a CONNECT target or a plain-HTTP host. It never decrypts TLS. The one TLS key it loads is a throwaway that the unit generates at each start, that signs nothing and that no agent trusts. It caches nothing.

  • Credentials. A request for a listed name must carry a per-agent proxy credential (HTTP Basic, the agent's name and its secret). swarm-controller mints one per agent at swarm/agents/<agent>/proxy-credential, and never replaces it. hive-c0re hashes the credentials of this hive's agents, and one more for the self-check, into services.hyperhive.deploy.hive-controller.squid.usersFile every 30 seconds. Squid's basic_ncsa_auth helper re-reads the file when it changes. A stopped agent stays in the file. Squid caches a good login for a minute. Nothing rotates a credential: to replace one, delete its leaf in the store and the next pass mints a new one. HTTP Basic crosses the bridge in clear between an agent and the proxy. See the credentials reference. A request without a valid credential gets a 407, once the destination checks below have passed it.

  • Agents and the proxy. An agent uses the proxy only when its hive says this host's squid is ready for it. hive-c0re puts two facts in every container's /etc/hyperhive-hive-facts, only on a hive that holds a store identity: HIVE_SQUID_PROXY (services.hyperhive.deploy.hive-controller.squid.listen, the bridge IP and port) and HIVE_EGRESS_NO_PROXY (the direct hosts below). A hive that predates them writes neither, and its agents go direct. hive-agent-proxy-credential.service (nix/agent-modules/egress-proxy.nix) runs in each agent's container, as the agent user, at start and every 10 minutes. With both facts present it logs in to the secret store with the agent's own certificate, reads swarm/agents/<agent>/proxy-credential and writes /run/hive-agent-proxy-credential/proxy.env (0600) with HTTPS_PROXY, HTTP_PROXY, their lower-case forms, NODE_USE_ENV_PROXY=1 and NO_PROXY/no_proxy. The proxy URL is http://<agent>:<credential>@<bridgeIp>:3128; Claude Code, Node, curl, git and reqwest send the userinfo as Proxy-Authorization. The harness, its bash tool, the subagents, the MCP servers and the pollers load the file with EnvironmentFile=- when they start. Until the facts and the credential both exist nothing writes the file, so those units have no proxy variable and go direct; a failed read keeps the file already in place. Nothing restarts on deploy: an agent takes the proxy at its next start, which is its next rebuild or restart. The credential is in the environment of every process those units start, so env in the agent's shell prints it. Login shells (nixos-container root-login, SSH) don't load the file.

  • Direct destinations. HIVE_EGRESS_NO_PROXY is the hosts this agent must reach without the proxy, and NO_PROXY is localhost,127.0.0.1,::1 followed by that list. hive-c0re computes it from the host's options: the bridge IP, the hive's domain (services.hyperhive.domain), the secret store and queue names (services.hyperhive.swarm.bao.domain, swarm.nats.domain), services.hyperhive.swarm.serviceDomains (the swarm UI and controller apex, the forge, matrix, attic, logs, metrics and otel names) and every name the gateway answers with the bridge IP (services.hyperhive.gateway.localNames). Squid refuses every destination that resolves to the bridge, and the hive's resolver answers the bridge address for those names, so they never go through the proxy. A bare name matches itself and its subdomains. Node's NO_PROXY has no CIDR support, so the list holds the bridge as an address.

  • Two lists. The credential decides the policy. hive-c0re writes a second file beside the users file, services.hyperhive.deploy.hive-controller.squid.unrestrictedFile (/var/lib/hive-squid/unrestricted, hive-core:squid, 0640): the names of the users whose agent holds the unrestricted_internet flag in the permissions the hive last read from the swarm, one per line, rewritten when it changes. The default is permissive: an agent with no record, or a record that never set the flag, holds it, the same default as the hyperhive.agent.unrestricted_internet metric. While the hive's read of the permissions is stale or failing, a new agent that already holds a credential counts as unrestricted. Squid reads the list when it parses its configuration, so squid-unrestricted-reload.path reloads squid when the file changes. A missing or empty file lists no one, and squid starts either way. A flag change reaches squid after the hive's next read of the permissions and its next 30-second pass. An older hive-c0re writes no list, so every credential on that hive is restricted. Agents must not start sending credentials until every hive runs the hive-c0re that writes it, or a legacy agent loses general internet.

  • Log mode. The operator can put an agent in log mode in the swarm UI. Its record keeps unrestricted_internet at false and adds internet_log: true, so a hive-c0re that doesn't know the field still enforces the list. For such an agent hive-c0re adds to agent-allow.conf, after the allow rules for every agent's own names, http_access allow !icpt <user> !allowed_dst log_mark and the same rule for the intercept ports, where log_mark is an annotate_transaction ACL that sets would_deny=1. The rules come last in the include, so every destination deny and deny !icpt !authed before them still applies, and a name the agent may reach never carries the mark. A request the rule lets through gets a second log line, would_deny user=<agent> method=<m> host=<h> port=<p> result=<r>, from the would_denied access_log in hive-squid.nix; the usual request line keeps its format. The would-deny tab of the egress dashboard in Grafana lists the hosts per agent. Only the operator changes the mode: nothing moves an agent to restricted after a time or a count. The self-check's users aren't agents, so deny-canary.invalid stays a 403.

  • Modes. An agent is in one of three modes, and the deny list (next bullet) applies in all of them:

    mode the deny list a name not on the allow list
    restricted blocked blocked
    log blocked allowed and logged
    unrestricted blocked allowed

  • Deny list. The swarm's deny list names destinations the proxy refuses whatever else allows them. The swarm controller keeps it as its own value (the deny key of the egress-allow bucket), each entry for every agent or for one, and serves it at GET/POST/DELETE /api/egress/denied-domains with the same body as the allow-list ({"domain": ".example.com", "agent": "atlas"}). Write a name as in the allow-list: at least two labels, one leading dot for the name and its subdomains. The controller refuses IP addresses, other wildcards, ports and paths. Unlike the allow-list it doesn't refuse the base list's names: denying one is the point. Every 30 seconds hive-c0re reads the list and renders services.hyperhive.deploy.hive-controller.squid.denyFile (/var/lib/hive-squid/deny.conf, hive-core:squid, 0640) for the agents it runs, and squid-unrestricted-reload.path reloads squid when it changes.

    • Precedence. The configuration includes the file after the destination denies and before deny !icpt !authed and every allow, so a denied name is refused even when the base list, the hive's allowedDomains, the swarm's allow-list, the unrestricted flag or log mode would allow it. A swarm-wide entry names no user, so every client gets a 403 for it, one without a credential too. An agent's entry also names the agent's squid user, so it binds that agent: a client that sent no credential gets the usual 407 for it. An agent's entry only adds to the swarm-wide ones; no entry of one scope lifts another.
    • Log line. A request a deny rule refused gets a second log line, egress_denied user=<agent or -> method=<m> host=<h> port=<p> result=<r>, in every mode. It differs from the would_deny line of log mode and from the TCP_DENIED/403 line every refusal has, so you can tell a deny-list hit from a name that isn't on the allow-list.
    • Who writes it. Only an operator's request to the controller writes the list. It has no automatic entries: nothing derives a deny from an agent's API-key environment or a linked account, and the stored type has no such field. An agent has no queue grant that reaches the key, and the controller's API sits behind the operator's SSO login: nginx checks it before it hands a request to the socket, which only nginx and root can open.
    • Empty list. With no entry deny.conf is empty and squid configures as it did without the list: the configuration gains one include line and the egress_denied log lines, and no rule. A hive on a hive-c0re that predates the list never reads the key and keeps its old configuration, so nothing on its wire changes.
    • Applied. Each hive reports the revision of the list it last rendered as egress_deny_applied in its status snapshot (see the hive snapshot). GET /api/egress/denied-domains answers the list's revision and behind, the hives that haven't reported it: an older hive-c0re never does, so it shows there rather than silently allowing the names. A hive that reads the list reports within about a minute. Reporting means the file holds the rules; squid reloads moments later.

  • Base list. services.hyperhive.deploy.hive-controller.squid.baseAllowedDomains is open to every authenticated user, whatever its flag: what the agent harness needs off the swarm. The default is the hosts Claude Code requires (https://code.claude.com/docs/en/network-config): api.anthropic.com for the model API and the usage probe (hive-agent/src/claude_usage_watch.rs), claude.ai for the account login, and platform.claude.com for the OAuth token exchange and refresh. An agent whose ANTHROPIC_BASE_URL points elsewhere needs no entry here: the swarm controller adds that host to the agent's own swarm entries (see Swarm allow-list below). The list omits the plugin marketplace clone from github.com, since a failed add isn't fatal. The rules below still refuse an entry that resolves to a private address.

  • Allow-list. services.hyperhive.deploy.hive-controller.squid.allowedDomains defaults to empty, plus the self-check canary (services.hyperhive.deploy.hive-controller.squid.selfcheck.canary), so the canary widens the restricted list by that one name. The swarm's own service names aren't on it: they resolve to the bridge or a private address, which the rules below refuse, and agents reach them directly. Each entry matches that exact name, or with a leading dot the name and its subdomains. A user on the unrestricted list reaches any other name too, on ports 80 and 443. Everyone else gets a 403 for it. allowedDomains is the hive's own list, always allowed whether or not the swarm has a list of its own (next bullet).

  • Swarm allow-list. The operator adds domains in the swarm UI (allowed domains); the swarm controller keeps them in one value of the egress-allow bucket, each for every agent or for one agent. Every 30 seconds hive-c0re reads it and writes two files beside the users file, services.hyperhive.deploy.hive-controller.squid.allowedFile (/var/lib/hive-squid/allowed) and services.hyperhive.deploy.hive-controller.squid.agentAllowFile (/var/lib/hive-squid/agent-allow.conf), hive-core:squid, 0640, and logs squid allow list: 2 swarm, 3 agent entries when they change. The swarm-wide names join the allow ACL, so a restricted user reaches them. A per-agent entry is a rule for that agent's squid user, http_access allow !icpt <user> <names>, and a second one for the user the intercept helper names (icpt <user> <names>), in the generated agent-allow.conf that the configuration includes at the point where the base list gets its allow: after every destination deny and deny !icpt !authed, so the rules under Denied before any allow still bind it, and before the unrestricted rule. Only this hive's agents get a rule. squid-unrestricted-reload.path reloads squid when either file changes, so there is no deploy or restart. A name is a DNS name of at least two labels. One leading dot (.example.com) also allows every subdomain, and is the only wildcard; the controller refuses IP addresses, other wildcards, ports, paths and the base list's names when you add the entry, and hive-c0re checks each name again and skips one that fails, or that the base list or the canary covers, or that another entry of the same list covers. A name that resolves to a private address stays a 403. The controller also adds entries itself, and only from the operator's own actions in the swarm UI. Saving an agent's API-key environment makes it take the host from ANTHROPIC_BASE_URL and from every variable named *_BASE_URL (the address an ACP agent that loads the same file reads), and keeps one derived entry for that agent per host. Each host passes the same check as an operator's entry, so an IP address, localhost and a base-list name get none. A save or a delete of the environment replaces the agent's derived entries, and destroying the agent drops all its entries. The agent can create its own backend-env leaf while the leaf is empty, so text found in the store never creates an entry: when the controller starts, it only drops derived entries whose exact host the stored environment no longer names, and logs a host it didn't derive. An environment saved before this existed gets its entries when the operator saves it once more. An agent on the hand-placed backendEnvironmentFile alone, and the opencode preset's acp.opencode.provider.baseUrl, get no derived entry, so add one in the swarm UI. A hive-c0re that predates the field renders a derived entry like any other. Linking a GitHub token, an external matrix account or an external forge account adds a derived entry for that agent per host it reaches, and deleting the account removes it: github.com, api.github.com, raw.githubusercontent.com, codeload.github.com and objects.githubusercontent.com for GitHub, and the host of the matrix account's homeserver or the forge account's url. Each entry's source (link:github, link:matrix:<account>, link:forge:<label>) keeps it apart from the environment's, which a save of the environment leaves alone; an entry without a source counts as the environment's. Only the controller writes these leaves, so when it starts it also adds the entries of accounts linked before this existed and drops those whose account no longer exists, and logs the counts (link egress: entries match the linked accounts). A store it can't read changes nothing. An agent.nix matrixAccounts homeserver gets no entry, since the controller never reads it; add one in the swarm UI. The swarm's list adds to the hive's allowedDomains: allowedFile holds those names plus the swarm-wide entries, so removing an entry never removes a name the hive lists itself. Until the controller has a list (it predates the feature, or nobody added an entry) allowedFile holds allowedDomains and agent-allow.conf is empty. A hive-c0re older than this rule replaces allowedDomains with the swarm's list once the controller has one, so every hive must run the new hive-c0re before the controller's first entry exists. A queue that stops answering for the bucket leaves a file the swarm wrote untouched, and a failed read leaves both files untouched. The first line of a file written from the swarm's list is a # source: swarm comment. The queue grant a hive needs to read the list comes with the controller's swarm-nats-auth; a hive that can't read it times out, keeps its files and warns after about 90 seconds.

  • Unlisted names. For a user who isn't on the unrestricted list, refused after the destination checks, which look the name up on the host.

  • Denied before any allow. Destinations that resolve to loopback, link-local, the bridge subnet, a private range (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16, fc00::/7), shared address space (100.64.0.0/10), 0.0.0.0/8, ::, multicast or reserved space, and IP-literal hosts, for every user including the unrestricted ones and for listed names. Squid resolves names on the host and connects from it, so without these a name pointing at 127.0.0.1, at the host's own LAN address or at a private neighbour would walk around the bridge→loopback DROP rule and the swarm's network. Squid checks the address it connects to, so a name that changes its answer between a check and the connect gains nothing. A listed name must therefore resolve to a public address on the host's resolver: a split-horizon name that answers a LAN address gets a 403, and so does the self-check canary, which is why it isn't a swarm service name. Only CONNECT to port 443 and plain HTTP to ports 80 and 443 pass.

  • Metrics. squid-exporter serves prometheus-squid-exporter on 127.0.0.1:9301. The hive collector scrapes it as job squid through services.hyperhive.otel.scrapeTargets. squid_up and squid_client_http_requests_total are the two to watch.

  • Logs. One syslog line per request reaches the journal as squid.service, and services.hyperhive.otel.journalUnits forwards that unit and the self-check's to the swarm's log store. A refusal shows as TCP_DENIED/403 with the client's source address. A request that a log-mode agent's rule let through also logs a would_deny line (see Log mode). A request a deny-list rule refused logs an egress_denied line (see Deny list).

  • Self-check. squid-selfcheck.timer runs squid-selfcheck.service every minute from the bridge IP, logging in as selfcheck.squid: CONNECT to services.hyperhive.deploy.hive-controller.squid.selfcheck.canary (default: cache.nixos.org, a public name that's also on the allow-list) must answer 200, to deny-canary.invalid 403, and to the canary without a credential 407. It then logs in as selfcheck-unrestricted.squid, which has the same secret and is always on the unrestricted list, and a CONNECT to services.hyperhive.deploy.hive-controller.squid.selfcheck.unrestrictedCanary (default: example.com, a name outside the allow-list) must answer 200. It writes one squid-selfcheck ok=… allow_status=… deny_status=… anon_status=… credential=present|absent unrestricted_canary=… unrestricted_status=… line per run and fails the unit when any is wrong. The password is the hive's swarm/hives/<hive>/squid-selfcheck secret, which squid-selfcheck-credential lands at /run/credstore/hive-squid.selfcheck. Before swarm-controller has minted it, or while hive-c0re hasn't written the users file, the line reads credential=absent or allow_status=407 and ok=false, and nothing else breaks. The host's resolver looks up the canary, so a host whose resolver maps that name to a loopback, bridge or private address fails the check.

Intercepted web traffic

A program that ignores the proxy variables still gets squid's policy for its web traffic. In the inet hive_egress table (nix/host-modules/hive-egress.nix) the intercept chain, at prerouting, redirects IPv4 TCP port 80 to <bridgeIp>:3129, an intercept listener of squid, and port 443 to <bridgeIp>:3131, the ClientHello relay (see ECH below), which sends the connection on to squid's other intercept listener, <bridgeIp>:3130. Squid reads the original destination from conntrack. The chain exists where the host runs squid (services.hyperhive.network.egress.intercept.enable).

  • Who gets redirected. Every flow from the bridge to port 80 or 443, except a flow to an address that stays on the bridge, to one of the host's own addresses or to the wg-hive interface, and a flow from a MAC in the set direct_mac. That set holds the MACs of the ports of every container that's not an agent (the swarm's service containers) and of the agents listed in services.hyperhive.deploy.hive-controller.squid.unrestrictedFile. So restricted and log agents get redirected, and unrestricted agents and service containers keep the direct path. Every pass of hive-priv egress-pin and every egress-load refills the set from the bindings it also uses for the bridge table, and hive-egress-direct.path runs a pass when hive-c0re rewrites the list; the timer covers the rest. A MAC in no binding is in no exemption: the flow gets redirected, and squid refuses it. A new agent that the list doesn't name yet counts as restricted until the next pass, and a flag change reaches the redirect after the hive's next read of the permissions and one pass. IPv6 and the other ports aren't redirected.

  • Who the flow belongs to. An intercepted request carries no proxy credential, so squid asks its external ACL hive_agent, which runs hive-priv egress-whois. The helper looks the source address up in dnsmasq's lease file for a MAC, the MAC in the bindings (/run/hive-egress-pins) for a machine, and takes the agent from the machine name h-<agent>. While antiSpoof is drop, an address on the bridge is the one leased to the MAC of the port it came from (see Agent bridge source check), so the source says which container sent the flow. Squid keeps an answer for 10 seconds and a refusal for 2. An address with no lease or an expired one, two MACs with the same address, a MAC no port holds, a machine that's no agent: none gets an answer, and squid refuses the flow with a TCP_DENIED/403 and user -. A host that has no bindings or leases yet refuses every intercepted flow rather than guess. With antiSpoof = log a container can forge the source, so use that mode to diagnose and not to run. The proxy port never takes this identity: a request without a credential through port 3128 still gets a 407, whatever its address.

  • Policy. The same lists as for the proxy port, matched on the agent the helper names: the destination denies, the deny list, the base list, the allow-list and the agent's own names, and log mode with its would_deny line. hive-c0re renders each agent rule twice, once for the credential (!icpt) and once for the helper's user (icpt).

  • HTTPS. Squid peeks at the ClientHello for the SNI and splices the connection. It never decrypts, and no agent trusts a CA of the host. Squid builds a CONNECT with the SNI as its host and applies every rule to it. A refused name gets a 403 that the client reads as a TLS error. Squid also refuses:

    • an SNI that doesn't resolve to the address the agent dialed (host_verify_strict; squid answers 409 and the client reads a TLS error). Squid resolves on the host's resolver, which is the same dnsmasq as the agent's, so a name with changing answers can fail now and then;
    • a ClientHello with no SNI, and a URL with an IP literal, which sends none: the host of the CONNECT stays an address, and the IP-literal denies refuse it;
    • bytes on port 80 that aren't an HTTP request. Bytes on port 443 that aren't a ClientHello never reach squid (see ECH below).

  • ECH. Squid 7.6 can't see the encrypted_client_hello extension, only the public name in the outer hello. An agent that holds an ECH configuration could send a hello with an allowed outer name, a destination address that name resolves to, and a hidden inner name that the same address serves. Port 443 therefore doesn't go to squid directly. hive-priv egress-ech-relay (hive-egress-ech-relay.service, egress.intercept.refuseEch, on by default) listens on <bridgeIp>:3131. It reads the first TLS records of each redirected connection, up to four records and 16 KiB of ClientHello within 10 seconds, and closes the connection when the ClientHello of a restricted agent carries the extension (type 0xfe0d, and the draft types 0xfe08 to 0xfe0c, 0xff02 and 0xffce). The relay only observes a log agent, as everywhere in log mode: its hello goes on like a clean one, and the relay logs it (see below). It closes a connection too, for any agent, when the bytes aren't a well-formed ClientHello, when they don't arrive in time, or when the connection wasn't redirected. It also closes a connection from a source that already has 256 open, and any connection beyond 2048 in all, so one agent can't use up the relay. Otherwise it connects to the destination the agent dialed and sends the bytes on, and squid sees an ordinary intercepted connection:

    • the relay connects from an address in 198.18.0.0/15, which the unit routes to lo, and an output rule DNATs its marked connections (mark 0x45434800) to squid's port 3130, so squid reads the original destination from conntrack as before;
    • the source address is 198.18.0.0 plus the low 17 bits of the agent's address, and egress-whois maps it back to the lease it came from, so the pool needs a bridge subnet of /15 or smaller (an assertion checks) and squid's log shows 198.18.x.y as the client address of an intercepted HTTPS flow. The user column still names the agent. The input chain drops a pool address that arrives from the bridge;
    • the relay logs to the journal of its unit, at most one line a second per source address, with the agent named the way squid names it (- when no lease and binding says). A refusal is hive-egress-ech-refused agent=<name> dst=<address>; an ECH hello of a log agent, which goes on, is hive-egress-ech-would-refuse agent=<name> dst=<address>; a hello that isn't well formed is hive-egress-hello-refused agent=<name> dst=<address> reason=<why>. Squid logs nothing for a refused flow;
    • the mode comes from agent-allow.conf, the include hive-c0re renders: an agent is in log mode when it holds the line acl agent_log_<name> proxy_auth <name>. Squid applies the same file, so the relay and squid can't disagree about an agent's mode, and the file changes when the operator changes the mode, with no second list to keep in step. A missing or unreadable file, or an agent with no lease, counts as restricted, and the relay refuses the hello;
    • while the relay is down the redirect reaches a closed port, so HTTPS from restricted and log agents fails. The unit restarts after 5 seconds.

    GREASE: a client with no ECH configuration, Chrome and Firefox by default, sends the extension anyway with random contents, and the protocol makes the two look the same on the wire. The relay can't tell them apart, so it treats both as ECH: a restricted agent's HTTPS fails when the client ignores the proxy variables, and a log agent's goes on with a would-refuse line. Through the proxy port (HTTPS_PROXY, a CONNECT) it works as before: that path has no interception. A host where that costs more than the gap does can set egress.intercept.refuseEch = false, which sends port 443 straight to squid and leaves ECH unchecked. The relay only sees the hello of a connection that the redirect catches. It doesn't touch UDP (QUIC), IPv6 or other ports.

  • Reading the log. Each intercepted HTTPS connection logs twice: first NONE_NONE/000 … CONNECT <address>:443 <agent>, before the ClientHello, then the CONNECT <name>:443 <agent> line with TCP_TUNNEL/200 or TCP_DENIED/403 (NONE_NONE/409 when the SNI and the address disagree). The client address of an intercepted HTTPS line is the relay's 198.18.x.y (see ECH above), and an intercepted HTTP request logs GET http://<name>/ <agent> with the agent's own address. A tunnel to an intercepted destination ends in ORIGINAL_DST/<address>. cache.log notes an ACL that met a flow squid refused before its request.

  • Rollout. The redirect, the relay, squid's configuration and hive-priv ship in one build. A hive that still runs an older build has no redirect, and its agents go direct as before. The agents that use the proxy variables see no change.

Container isolation

Each agent container runs in a private network namespace with a dedicated veth pair attached to the bridge. The following table summarises what the nix side sets up unconditionally:

effect mechanism
IP forwarding boot.kernel.sysctl."net.ipv4.ip_forward" = 1
Internet NAT networking.nat { enable = true; internalInterfaces = [ bridgeName ]; } — MASQUERADE on packets leaving via any external NIC
Loopback DROP nix/host-modules/hive-egress.nix — own nftables table inet hive_egress (whichever firewall the host runs): drops bridge → 127.0.0.0/8 and ::1 traffic; the kernel default route_localnet=0 is the first layer
Forward log nix/host-modules/hive-egress.nix — the same table's forward chain logs and counts what the host routes from the bridge to the uplink, and refuses it with egress.forward = "reject" (see Agent bridge egress log below)
Source check nix/host-modules/hive-egress.nix — the bridge hive_egress table drops a frame whose source MAC or address differs from the sending container's own (see Agent bridge source check below)
Gateway access networking.firewall.interfaces.<bridge>.allowedTCPPorts = [ 80 443 ] — lets isolated agents (private netns, veth on bridge) reach nginx on the host
c0re signal HIVE_NETWORK_BRIDGE, HIVE_NETWORK_SUBNET in systemd.services.hive-c0re.environment — both required; hive-c0re refuses to start without them

HIVE_NETWORK_SUBNET is the host-side bridge IP + prefix (e.g. 10.42.0.1/24), not the canonical network address. The Rust side must normalise (bitwise-AND with mask) before subnet membership checks or address arithmetic.

What the Rust side does

hive-c0re reads HIVE_NETWORK_BRIDGE + HIVE_NETWORK_SUBNET and passes PRIVATE_NETWORK=1, LOCAL_ADDRESS= (empty), HOST_ADDRESS=<bridge-ip>, and HOST_BRIDGE=<bridgeName> when creating or updating containers: lifecycle::set_nspawn_flags hands them to hive-priv, which rewrites /etc/nixos-containers/<container>.conf. hive-c0re validates both variables once at daemon startup, not per container: they're process-global, so a missing or malformed value is a misconfigured daemon rather than one bad container, and failing at boot gives a single diagnostic instead of one per agent. No non-isolated mode exists to fall back to. hive-c0re leaves LOCAL_ADDRESS empty so the container's dhcpcd acquires an address from the bridge dnsmasq pool (networking.useDHCP = true in nix/agent-modules/network.nix). This applies to every agent container, manager included.

HOST_ADDRESS is the bridge gateway IP (the address part of HIVE_NETWORK_SUBNET, via lifecycle::bridge_gateway_ip — taken verbatim so a non-.1 operator override still resolves to wherever the bridge actually lives). It's load-bearing: nixos-container's container-side network setup only installs a default route (ip route add default via $HOST_ADDRESS) when HOST_ADDRESS is non-empty. In bridge mode nixos-container skips the host-side address/route setup, so writing it only affects the container's default route — without it the container comes up with an IP but no path off the bridge subnet (no internet, no api.anthropic.com).

How the isolated container gets its resolver

nixos-container copies the host's /etc/resolv.conf into the container at every start. The host resolver (for example 127.0.0.53 from systemd-resolved, or a LAN router) is unreachable from a private netns and isn't authoritative for the hive's own zones, so it's replaced with the bridge dnsmasq at boot. Because the copy happens on every start, it would clobber a declarative environment.etc."resolv.conf" — so the wiring is runtime:

  • hive-priv drops a marker file (/etc/hyperhive-bridge-dns, carrying the gateway IP) into each container's /etc.
  • the hyperhive-isolated-dns oneshot (nix/agent-modules/network.nix), gated on that marker, rewrites /etc/resolv.conf to nameserver <gateway-ip> at boot. It's ordered before the harness (hive-ag3nt), the matrix daemon, and the forge-token fetch so the resolver is correct before the first DNS lookup.

Why isolation is safe: hive-c0re's control-plane sockets are unix domain sockets bind-mounted into containers, not network listeners — see the Control plane (no network) bullet under Network map above. PRIVATE_NETWORK=1 has no effect on a path that never touches the network stack.

The nix side also enables IP forwarding + NAT (agents reach the internet through the host) and drops bridge-subnet → loopback traffic from its own nftables table (inet hive_egress, nix/host-modules/hive-egress.nix, which also holds the bridge hive_egress table of Agent bridge source check; defence-in-depth against a compromised agent reaching the c0re dashboard HTTP at 127.0.0.1). Agents have no legitimate reason to reach the dashboard over loopback — the hive-c0re admin socket is a UDS, not TCP.

Agent bridge egress log

services.hyperhive.network.egress.forward sets what the forward chain of the inet hive_egress table does with traffic the host routes from the agent bridge to anywhere but the bridge and the wg-hive mesh interface. It takes accept, log (the default) or reject. In log mode the chain blocks nothing. It logs each new flow and counts every packet. In reject mode it also refuses the flow.

  • Who is exempt. In log and reject mode alike, the chain accepts the mesh traffic first, then the traffic of every MAC in the direct_mac set: the unrestricted agents and the service containers, which the intercept chain leaves alone too (see Intercepted web traffic). They leave no line, so a hive-egress-would-drop line is exactly a flow that reject would refuse, and a service container keeps every port and protocol under reject.
  • What reject refuses. Everything from the bridge to the uplink that the chain didn't exempt, whether a flow is new or not: TCP on any port but the 80 and 443 that squid takes, UDP (DNS to an outside resolver, QUIC and NTP included), ICMP and IPv6. A TCP flow gets a reset, so the connect fails at once instead of stalling. Everything else gets ICMP administratively prohibited. The log lines of this mode carry the prefix hive-egress-reject (and hive-egress-reject-unmetered ); the rate limit and the counter work as in log mode. Still open to every agent are the host's own addresses: the proxy ports, DNS and DHCP from the bridge address, and the gateway.
  • Switching on. Leave log running for about a week, then read the hive-egress-would-drop lines. Each is a flow that reject would refuse; a flow that belongs to a legitimate use needs a fix first (a service container missing from direct_mac, an agent that should be unrestricted). Then set services.hyperhive.network.egress.forward = "reject". Making it the default is a one-line change of the option's default.
  • What reject can break.
    • A service container with no pin yet isn't in direct_mac, so the chain refuses it until hive-egress-pin binds its port. The bridge table drops its frames until then anyway.
    • A change of an agent's unrestricted flag reaches direct_mac within 30 seconds at most, on the timer of hive-egress-pin. The chain refuses the agent in the gap.
    • An unreadable unrestricted file lists no agent, so unrestricted agents lose every non-web flow.
    • An agent in per-agent log mode keeps its web traffic, which squid only observes, and loses every other port and protocol.
  • Which traffic. IPv4 and IPv6 alike (the table is inet). The chain never sees traffic to the host's own addresses, because the host delivers it locally: a connection to squid on <bridgeIp>:3128, to dnsmasq or to the gateway leaves no line. Nor does anything squid or the host opens itself. The web traffic of a restricted or log agent that the intercept chain redirects into squid reaches the host too, so squid's log shows it instead (see Intercepted web traffic).
  • The log line. The kernel log carries one line per new flow, with the prefix hive-egress-would-drop followed by the usual netfilter fields (IN, OUT, SRC, DST, PROTO and the ports). Read them with journalctl -k -g hive-egress-would-drop.
  • Rate limit. Each source address gets a burst of 20 lines, then one line a second, and the chain forgets it after a minute without a new flow. A flood from one agent can't hide another's flows, and it adds about 250 bytes of journal a second per source address at most. A flow the per-source limit can't log (the address is over its rate, or the table of addresses is full) gets one line with the prefix hive-egress-would-drop-unmetered instead, from a limit shared by all addresses: 5 lines a second after a burst of 50, about 100 MB of journal a day at most. A flow never gets both lines. The counter in nft list chain inet hive_egress forward counts every packet, logged or not.
  • Which agent. SRC is the agent's address on the bridge. Containers get theirs from the hive's dnsmasq, which records each lease in /var/lib/dnsmasq/dnsmasq.leases, one line per lease: <expiry> <mac> <ip> <hostname> <client-id>. Look the SRC address up in the third column; the hostname names the container. A lease lasts one hour, so look an address up soon after the line appears.
  • Trust. A frame that carries another container's source address never reaches the forward chain while antiSpoof is drop (see Agent bridge source check), so SRC names the container that sent the flow. In log mode a container can still forge it, so a log from a host that ran in log mode during the soak doesn't show who sent a flow.

Agent bridge source check

Every container on the agent bridge is root in its own network namespace and shares one L2 segment with the others, so it can send with another container's MAC or address. services.hyperhive.network.egress.antiSpoof sets what the bridge hive_egress table (nix/host-modules/hive-egress.nix) does about it. It takes drop (the default) or log. Two tables, not one: the bridge carries traffic between two ports, and the inet hooks never see it as coming from a port. Only a bridge-family hook at prerouting sees which veth a frame came in on, so the source check lives in that table and the host-side filtering stays in inet hive_egress.

  • What a port may send. The frame's MAC is the one bound to the port. Then, by protocol: IPv4 from an address leased to that MAC, or DHCP from 0.0.0.0 to 255.255.255.255; ARP whose sender is the port's MAC and an address leased to it, or an ARP probe from 0.0.0.0; IPv6 from a link-local address (and from :: in neighbour solicitations and listener reports, which precede an address). A message to UDP port 67 must carry the port's own MAC as its chaddr, so a container can't lease, renew or release an address for another. The table refuses a VLAN tag, and everything else counts as a spoof.
  • A port nobody owns sends nothing. Not even DHCP. A container's port is bound a moment after its veth appears, and the container's DHCP client retries until then.
  • Where the bindings come from. Addresses are dynamic, so the table's sets fill at run time. hive-priv egress-pin binds a port to the MAC of the link at the other end of its veth pair, found through the container systemd registers for it. udev starts it for each new vb-* port (hive-egress-pin@<port>.service), and a timer (hive-egress-pin.timer, every 30 seconds) runs it for the ports that still have no binding. The timer reads no container while every port has one. The hive-egress unit loads the tables with hive-priv egress-load: it looks up the bindings first, while the old tables still filter, and puts them and the leases from dnsmasq's lease file into the same nft transaction as the tables. A start or reload never leaves the bridge with empty sets. If loading the leases fails, the unit logs a warning and loads the tables with the bindings alone. A binding stays with the port until the port goes: root in the container can't move the port to another MAC afterwards (state in /run/hive-egress-pins). A port never takes a MAC that a binding already holds, so a newcomer can't evict an established container. If two new ports claim one MAC, neither gets a binding. dnsmasq's dhcp-script (hive-priv egress-lease) adds each lease as a (MAC, address) pair that times out with the lease.
  • The window. The binding takes the MAC the container's link has when the command first looks, normally within a second of the veth appearing. Until then root in the container could change its own MAC, but in that time only the container's early init runs, which is nix-built, and the agent's code starts later. A container that did change its MAC first gets a binding for that MAC only if nobody holds it; a lease follows the chaddr the port may send. A port refused because it claimed a held MAC sends nothing, and the timer binds it once its link has a MAC of its own.
  • The log. A refused frame (all of them with drop, the same ones with log) gets a kernel log line with the prefix hive-egress-spoof , rate limited per port: a burst of 20, then one a second. A frame its port can't log gets a line with the prefix hive-egress-spoof-unmetered instead, from a limit shared by all ports: 5 lines a second after a burst of 50. Read them with journalctl -k -g hive-egress-spoof. IN= is the port, vb-<machine> (vb-h-<agent> for an agent), and SRC and MAC=<dst>:<src>:<type> are what the container sent; the port names the sender, because a bound port sends only its own MAC. The counters in nft list chain bridge hive_egress spoof count every refused frame.
  • Why drop is the default. A container that sends from its leased address and its own MAC is never refused, so drop breaks nothing a container does legitimately, and without it the egress log can't tell agents apart. log is for a live hive where the table refuses a legitimate container (a binding that failed shows as hive-egress-pin errors in the journal and IN= lines for that port): the same lines and counters, nothing blocked.
  • What it doesn't cover. A container can still send from any IPv6 link-local address and answer DHCP as a server on its own leased address.

Enforcement comes with later changes in the same series: the log only shows what such a change would cut.

Cross-references

  • docs/4-implementation/networking/gateway.md — vhost map + the gateway's other duties