The collector's oidc/swarm authenticator verifies the bearer offline as a JWT. The controller's authelia client was registered without accessTokenSignedResponseAlg, so authelia mints it an opaque authelia_at_ handle, which that authenticator cannot verify: by config, no push from the vcs or jobq exporter can be accepted. Register it with RS256 like every other client whose token lands on an oidc/* receiver. The SDK logs a failed export at debug only, so a refused push never reached the journal. AuthenticatedHttpClient now warns on a failed token mint, a transport error, or a non-2xx answer (status plus a body snippet), once per kind of failure and again every 10 minutes with a count, and logs once when a push succeeds after failing. Refs #5393
52 lines
2.3 KiB
Nix
52 lines
2.3 KiB
Nix
# Glue: register the swarm controller as an OIDC client wherever authelia runs.
|
|
#
|
|
# ONE PAIRING PER FILE — swarm-controller ← authelia, and nothing else. Deleting
|
|
# this leaves a controller whose token requests authelia answers with a 401, and
|
|
# no minted secret for swarm-secret-publish to put in the store.
|
|
#
|
|
# ⚠️ Gated on authelia being HERE, and NOT on this host running the controller.
|
|
# A client is a row in THIS host's provider config, so it is declared where that
|
|
# config is rendered; ./swarm-controller.nix's `config` hangs off
|
|
# `deploy.swarm-controller.enable`, which answers a question about a different
|
|
# host. Same shape as ./glue-grafana-oidc-client.nix.
|
|
#
|
|
# Registered whether or not the swarm runs a controller: nothing in `swarm.*`
|
|
# records that, so the cost of a swarm without one is one unused client and one
|
|
# unused minted secret.
|
|
{
|
|
lib,
|
|
config,
|
|
...
|
|
}:
|
|
let
|
|
hyperhiveCfg = config.services.hyperhive;
|
|
deployCfg = hyperhiveCfg.deploy;
|
|
queueClientId = hyperhiveCfg.swarm.controller.queueClientId;
|
|
in
|
|
{
|
|
config = lib.mkIf deployCfg.authelia.enable {
|
|
# One declaration, two readers: the id is the read-only option the
|
|
# controller authenticates under, so a mismatch cannot be introduced here.
|
|
services.hyperhive.swarm.authelia.oidc.clients = [
|
|
{
|
|
id = queueClientId;
|
|
description = "HyperHive swarm controller";
|
|
# `client_credentials`: a daemon authenticating as itself, with nobody
|
|
# to redirect. Declared rather than inferred from an empty redirect
|
|
# list, because authelia permits only the grants a client names and an
|
|
# omitted `grant_types` means authorization-code alone.
|
|
kind = "machine";
|
|
# Its own id as its own permitted audience. The controller mints a
|
|
# token for `SWARM_CONTROLLER_OTEL_AUDIENCE` (./glue-controller-env.nix),
|
|
# and ./swarm-otel.nix's `oidc/swarm` authenticator checks for this
|
|
# same id; either half without the other is a 401 several layers from
|
|
# this line.
|
|
audience = [ queueClientId ];
|
|
# `oidc/swarm` verifies the token offline against `/jwks.json`, which
|
|
# authelia's default opaque handle fails. The queue and the authelia
|
|
# bridge introspect, which reads a JWT just as well.
|
|
accessTokenSignedResponseAlg = "RS256";
|
|
}
|
|
];
|
|
};
|
|
}
|