Files
netbird/shared/management/proto/proxy_service.proto
Maycon Santos fa1e241aea [management, client, proxy] Follow-up fixes for private reverse-proxy services (#6268)
* fix(proxy): gate tunnel-peer fast-path on inbound listener marker

forwardWithTunnelPeer previously accepted any RFC1918 / ULA / CGNAT
source IP, so a public client whose address happened to fall in those
ranges could bypass the configured operator auth scheme by colliding
with a known tunnel IP. The fast-path is now gated on
TunnelLookupFromContext(r.Context()) being present — that context value
is attached only by the per-account inbound (overlay) listener, so the
host-facing listener never enters this branch.

Tests updated to reflect the new requirement: requests that don't
carry the inbound marker now fall through to the regular auth flow.

* fix(proxy): harden inbound listener resource + startup-ctx handling

Three correctness fixes on the per-account inbound path, with tests:

- Close the logrus ErrorLog PipeWriter on tearDown. WriterLevel hands
  back an *io.PipeWriter backed by a pipe + scanner goroutine that the
  caller owns; the two writers per account (https + plain) were never
  closed, leaking the pipe and goroutine on every teardown.
- Run the post-Start hooks on context.Background(). runClientStartup
  is launched in a goroutine from AddPeer and was inheriting the
  caller's request-scoped ctx, so a cancelled request could abort the
  inbound bring-up or fail the management status notification. The
  tail is split into notifyClientReady so the contract is testable.

Tests cover the PipeWriter close behaviour and assert the readyHandler
+ NotifyStatus calls receive a non-cancelled background context.

* feat(proxy): short-circuit peer-own-target loops with 421

When a peer that hosts the target of a private service dials its own
service URL the request was being looped through the proxy and back
over WireGuard to the same peer — twice the WG round-trip for no
benefit, with no signal to the caller that something was wrong.

Add isSelfTargetLoop to ReverseProxy.ServeHTTP: when the request
arrived on the per-account overlay listener (IsOverlayOrigin) and the
source tunnel IP matches the target host, refuse the request with 421
Misdirected Request and a body pointing the operator at the backend
directly.

The gate is scoped to overlay origin so requests on the public
listener that happen to share a source IP with the target host are
forwarded normally.

* fix(management): private-service validation + tunnel-IP lookup semantics

- Require an explicit port for L4 cluster targets. validateL4Target
  exempted TargetTypeCluster from the port check, but buildPathMappings
  serializes every L4 target via net.JoinHostPort(host, port) — port=0
  shipped a ":0" upstream. Cluster targets use the same Host/Port
  fields, so the same requirement applies.
- GetPeerByIP returns NotFound on a tunnel-IP miss instead of mapping
  every error to Internal. The proxy's ValidateTunnelPeer probes IPs
  that legitimately aren't in the roster; the miss is expected and now
  distinguishable from a real store failure.
- Thread ctx into getClusterCapability's gorm query so a cancelled
  request doesn't keep the store busy.

Tests updated for the L4-cluster port requirement and the GetPeerByIP
NotFound path.

* fix(client): include offlinePeers in PeerStateByIP lookup

ReplaceOfflinePeers moves peers into d.offlinePeers but PeerStateByIP
only scanned d.peers. Callers (the local DNS filter via
localPeerConnectivity, embed.Client.IdentityForIP used by the
proxy's tunnel-peer validator) were treating known-but-offline peers
as unknown, which:

- causes the DNS filter to keep returning records pointing at peers
  that have no live tunnel, AND
- makes the proxy's local-roster check deny a request from such a
  peer rather than letting the cached management RPC carry the
  authorisation decision.

Search both slices in PeerStateByIP. Adds a unit test for the IPv4
and IPv6 offline-match paths.

* fix(rest): reject empty Delete path params in reverse-proxy clients

ReverseProxyClustersAPI.Delete and ReverseProxyTokensAPI.Delete passed
the path parameter into url.PathEscape without an empty check.
PathEscape("") returns "" which collapses the request onto the
collection endpoint ("/api/reverse-proxies/clusters/" /
"/api/reverse-proxies/proxy-tokens/"), so a caller bug delete with no
id reached a routable URL with surprising semantics (typically 405).

Short-circuit with a typed error before the request is built. Tests
mount a handler on the collection path that fails the test if hit, so
the regression is impossible to reintroduce silently.

* chore(api,ci,docs,test): private-service schema, proto-check, fixups

Non-functional cleanups and contract/CI hardening around the
private-service work:

API schema (openapi.yml):
- Require a non-empty access_groups and mode=http when private=true,
  on both Service and ServiceRequest, mirroring
  validatePrivateRequirements. mode stays optional-but-constrained
  (empty defaults to http server-side), matching runtime.

CI (proto-version-check.yml):
- Cover renamed .pb.go files (read base via previous_filename).
- Match protoc-gen-go-grpc version headers (optional "- " prefix and
  -gen-go-grpc suffix) so grpc-generated files are in scope.

Docs / comments:
- Reword Config field docs to say defaults are applied at Server.Start
  (initDefaults), not New.
- Rename the obsolete --private-inbound flag to --private across
  comments and the proto doc.

Pre-existing test fixups surfaced by review:
- Repair the integration-tagged validate_session_test.go (SignToken
  signature growth + new Manager interface methods).
- Fix the CI-skip boolean precedence so Windows isn't skipped
  unconditionally.
- Guard the router.HTTPListener type assertion with comma-ok.

* fix(proxy): background ctx for already-started AddPeer notification

The earlier ctx fix covered the async runClientStartup path but missed
the synchronous branch: when a service is added to an already-started
client, AddPeer called NotifyStatus with the caller's request-scoped
ctx. A cancelled request/stream could drop the connected notification
to management. Use context.Background() here too, matching
notifyClientReady.

Extends TestNetBird_AddPeer_ExistingStartedClient_NotifiesStatus to
pass a pre-cancelled caller ctx and assert the notification still ran
on a non-cancelled context.

* use the cmd context for roundtripper
2026-06-02 13:40:09 +02:00

379 lines
13 KiB
Protocol Buffer

syntax = "proto3";
package management;
option go_package = "/proto";
import "google/protobuf/duration.proto";
import "google/protobuf/timestamp.proto";
// ProxyService - Management is the SERVER, Proxy is the CLIENT
// Proxy initiates connection to management
service ProxyService {
rpc GetMappingUpdate(GetMappingUpdateRequest) returns (stream GetMappingUpdateResponse);
// SyncMappings is a bidirectional stream that replaces GetMappingUpdate for
// new proxies. The proxy sends an initial SyncMappingsRequest to start the
// stream and then sends an ack after each batch is fully processed.
// Management waits for the ack before sending the next batch, providing
// application-level back-pressure during large initial syncs.
// Old proxies continue using GetMappingUpdate; old management servers
// return Unimplemented for this RPC and proxies fall back.
rpc SyncMappings(stream SyncMappingsRequest) returns (stream SyncMappingsResponse);
rpc SendAccessLog(SendAccessLogRequest) returns (SendAccessLogResponse);
rpc Authenticate(AuthenticateRequest) returns (AuthenticateResponse);
rpc SendStatusUpdate(SendStatusUpdateRequest) returns (SendStatusUpdateResponse);
rpc CreateProxyPeer(CreateProxyPeerRequest) returns (CreateProxyPeerResponse);
rpc GetOIDCURL(GetOIDCURLRequest) returns (GetOIDCURLResponse);
// ValidateSession validates a session token and checks user access permissions.
// Called by the proxy after receiving a session token from OIDC callback.
rpc ValidateSession(ValidateSessionRequest) returns (ValidateSessionResponse);
// ValidateTunnelPeer resolves an inbound peer by its WireGuard tunnel IP and
// checks the resolved user's access against the service's access_groups.
// Acts as a fast-path equivalent of OIDC for requests originating on the
// netbird mesh: when the source IP maps to a known peer in the calling
// account and that peer is in the service's access_groups, the proxy can
// issue a session cookie without redirecting through the OIDC flow.
// Mirrors ValidateSession's response shape.
rpc ValidateTunnelPeer(ValidateTunnelPeerRequest) returns (ValidateTunnelPeerResponse);
}
// ProxyCapabilities describes what a proxy can handle.
message ProxyCapabilities {
// Whether the proxy can bind arbitrary ports for TCP/UDP/TLS services.
optional bool supports_custom_ports = 1;
// Whether the proxy requires a subdomain label in front of its cluster domain.
// When true, accounts cannot use the cluster domain bare.
optional bool require_subdomain = 2;
// Whether the proxy has CrowdSec configured and can enforce IP reputation checks.
optional bool supports_crowdsec = 3;
// Whether the proxy is running embedded in the netbird client and serving
// exclusively over the WireGuard tunnel (i.e. `netbird proxy` rather than
// the standalone netbird-proxy binary). Surfaces upstream so dashboards can
// distinguish per-peer / private clusters from centralised ones.
optional bool private = 4;
// Whether the proxy enforces ProxyMapping.private (fails closed on ValidateTunnelPeer failure). Management MUST NOT stream private mappings to proxies that don't claim this.
optional bool supports_private_service = 5;
}
// GetMappingUpdateRequest is sent to initialise a mapping stream.
message GetMappingUpdateRequest {
string proxy_id = 1;
string version = 2;
google.protobuf.Timestamp started_at = 3;
string address = 4;
ProxyCapabilities capabilities = 5;
}
// GetMappingUpdateResponse contains zero or more ProxyMappings.
// No mappings may be sent to test the liveness of the Proxy.
// Mappings that are sent should be interpreted by the Proxy appropriately.
message GetMappingUpdateResponse {
repeated ProxyMapping mapping = 1;
// initial_sync_complete is set on the last message of the initial snapshot.
// The proxy uses this to signal that startup is complete.
bool initial_sync_complete = 2;
}
enum ProxyMappingUpdateType {
UPDATE_TYPE_CREATED = 0;
UPDATE_TYPE_MODIFIED = 1;
UPDATE_TYPE_REMOVED = 2;
}
enum PathRewriteMode {
PATH_REWRITE_DEFAULT = 0;
PATH_REWRITE_PRESERVE = 1;
}
message PathTargetOptions {
bool skip_tls_verify = 1;
google.protobuf.Duration request_timeout = 2;
PathRewriteMode path_rewrite = 3;
map<string, string> custom_headers = 4;
// Send PROXY protocol v2 header to this backend.
bool proxy_protocol = 5;
// Idle timeout before a UDP session is reaped.
google.protobuf.Duration session_idle_timeout = 6;
// When true, the proxy dials this target via the host's network stack
// instead of through the embedded NetBird client. Useful for upstreams
// reachable without WireGuard (public APIs, LAN services, localhost
// sidecars). Defaults to false — embedded client is the standard path.
bool direct_upstream = 7;
}
message PathMapping {
string path = 1;
string target = 2;
PathTargetOptions options = 3;
}
message HeaderAuth {
// Header name to check, e.g. "Authorization", "X-API-Key".
string header = 1;
// argon2id hash of the expected full header value.
string hashed_value = 2;
}
message Authentication {
string session_key = 1;
int64 max_session_age_seconds = 2;
bool password = 3;
bool pin = 4;
bool oidc = 5;
repeated HeaderAuth header_auths = 6;
}
message AccessRestrictions {
repeated string allowed_cidrs = 1;
repeated string blocked_cidrs = 2;
repeated string allowed_countries = 3;
repeated string blocked_countries = 4;
// CrowdSec IP reputation mode: "", "off", "enforce", or "observe".
string crowdsec_mode = 5;
}
message ProxyMapping {
ProxyMappingUpdateType type = 1;
string id = 2;
string account_id = 3;
string domain = 4;
repeated PathMapping path = 5;
string auth_token = 6;
Authentication auth = 7;
// When true, the original Host header from the client request is passed
// through to the backend instead of being rewritten to the backend's address.
bool pass_host_header = 8;
// When true, Location headers in backend responses are rewritten to replace
// the backend address with the public-facing domain.
bool rewrite_redirects = 9;
// Service mode: "http", "tcp", "udp", or "tls".
string mode = 10;
// For L4/TLS: the port the proxy listens on.
int32 listen_port = 11;
AccessRestrictions access_restrictions = 12;
// NetBird-only: the proxy MUST call ValidateTunnelPeer and fail closed; operator auth schemes are bypassed.
bool private = 13;
}
// SendAccessLogRequest consists of one or more AccessLogs from a Proxy.
message SendAccessLogRequest {
AccessLog log = 1;
}
// SendAccessLogResponse is intentionally empty to allow for future expansion.
message SendAccessLogResponse {}
message AccessLog {
google.protobuf.Timestamp timestamp = 1;
string log_id = 2;
string account_id = 3;
string service_id = 4;
string host = 5;
string path = 6;
int64 duration_ms = 7;
string method = 8;
int32 response_code = 9;
string source_ip = 10;
string auth_mechanism = 11;
string user_id = 12;
bool auth_success = 13;
int64 bytes_upload = 14;
int64 bytes_download = 15;
string protocol = 16;
// Extra key-value metadata for the access log entry (e.g. crowdsec_verdict, scenario).
map<string, string> metadata = 17;
}
message AuthenticateRequest {
string id = 1;
string account_id = 2;
oneof request {
PasswordRequest password = 3;
PinRequest pin = 4;
HeaderAuthRequest header_auth = 5;
}
}
message HeaderAuthRequest {
string header_value = 1;
string header_name = 2;
}
message PasswordRequest {
string password = 1;
}
message PinRequest {
string pin = 1;
}
message AuthenticateResponse {
bool success = 1;
string session_token = 2;
}
enum ProxyStatus {
PROXY_STATUS_PENDING = 0;
PROXY_STATUS_ACTIVE = 1;
PROXY_STATUS_TUNNEL_NOT_CREATED = 2;
PROXY_STATUS_CERTIFICATE_PENDING = 3;
PROXY_STATUS_CERTIFICATE_FAILED = 4;
PROXY_STATUS_ERROR = 5;
}
// SendStatusUpdateRequest is sent by the proxy to update its status
message SendStatusUpdateRequest {
string service_id = 1;
string account_id = 2;
ProxyStatus status = 3;
bool certificate_issued = 4;
optional string error_message = 5;
// Per-account inbound listener state for the account that owns
// service_id. Populated only when --private is enabled and the
// embedded client for the account is up. Field numbers >=50 reserved
// for observability extensions.
optional ProxyInboundListener inbound_listener = 50;
}
// ProxyInboundListener describes a per-account inbound listener that the
// proxy has bound on the embedded netstack of the account's WireGuard
// client. Surfaced so dashboards can render "this account is reachable
// at <tunnel_ip>:<https_port> on this proxy".
message ProxyInboundListener {
// Tunnel IP the embedded netstack listens on. Same address other peers
// in the account see for the proxy peer.
string tunnel_ip = 1;
// TLS port served on tunnel_ip (auto-detected, default 443).
uint32 https_port = 2;
// Plain-HTTP port served on tunnel_ip (auto-detected, default 80).
uint32 http_port = 3;
}
// SendStatusUpdateResponse is intentionally empty to allow for future expansion
message SendStatusUpdateResponse {}
// CreateProxyPeerRequest is sent by the proxy to create a peer connection
// The token is a one-time authentication token sent via ProxyMapping
message CreateProxyPeerRequest {
string service_id = 1;
string account_id = 2;
string token = 3;
string wireguard_public_key = 4;
string cluster = 5;
}
// CreateProxyPeerResponse contains the result of peer creation
message CreateProxyPeerResponse {
bool success = 1;
optional string error_message = 2;
}
message GetOIDCURLRequest {
string id = 1;
string account_id = 2;
string redirect_url = 3;
}
message GetOIDCURLResponse {
string url = 1;
}
message ValidateSessionRequest {
string domain = 1;
string session_token = 2;
}
message ValidateSessionResponse {
bool valid = 1;
string user_id = 2;
string user_email = 3;
string denied_reason = 4;
// peer_group_ids carries the calling user's group memberships so the
// proxy can authorise policy-aware middlewares without an additional
// management round-trip.
repeated string peer_group_ids = 5;
// peer_group_names carries the human-readable display names for the
// ids in peer_group_ids, ordered identically (positional pairing).
// Stamped onto upstream requests as X-NetBird-Groups so downstream
// services can read names rather than opaque ids.
repeated string peer_group_names = 6;
}
// ValidateTunnelPeerRequest carries the inbound peer's tunnel IP and the
// service domain whose group requirements should gate access. The calling
// account is inferred from the proxy's gRPC metadata (ProxyToken).
message ValidateTunnelPeerRequest {
string tunnel_ip = 1;
string domain = 2;
}
// ValidateTunnelPeerResponse mirrors ValidateSessionResponse plus a freshly
// minted session_token: when valid is true, the proxy installs the token as
// a session cookie so subsequent requests skip the management round-trip,
// matching the OIDC flow's UX. denied_reason values:
// "peer_not_found" — no peer with that tunnel IP in the calling account
// "no_user" — peer exists but is not bound to a user
// "service_not_found"
// "account_mismatch"
// "not_in_group" — peer resolved but not in service.access_groups
message ValidateTunnelPeerResponse {
bool valid = 1;
string user_id = 2;
string user_email = 3;
string denied_reason = 4;
// session_token is set only when valid is true. Same shape as the JWT
// the OIDC flow produces — proxy installs it via setSessionCookie so the
// tunnel fast-path is indistinguishable from OIDC for subsequent requests.
string session_token = 5;
// peer_group_ids carries the resolved peer's user group memberships so
// the proxy can authorise policy-aware middlewares without an additional
// management round-trip.
repeated string peer_group_ids = 6;
// peer_group_names carries the human-readable display names for the
// ids in peer_group_ids, ordered identically (positional pairing).
// Stamped onto upstream requests as X-NetBird-Groups so downstream
// services can read names rather than opaque ids.
repeated string peer_group_names = 7;
}
// SyncMappingsRequest is sent by the proxy on the bidirectional SyncMappings
// stream. The first message MUST be an init; all subsequent messages MUST be
// acks.
message SyncMappingsRequest {
oneof msg {
SyncMappingsInit init = 1;
SyncMappingsAck ack = 2;
}
}
// SyncMappingsInit is the first message on the stream, carrying the same
// identification fields as GetMappingUpdateRequest.
message SyncMappingsInit {
string proxy_id = 1;
string version = 2;
google.protobuf.Timestamp started_at = 3;
string address = 4;
ProxyCapabilities capabilities = 5;
}
// SyncMappingsAck is sent by the proxy after it has fully processed a batch.
// Management waits for this before sending the next batch.
message SyncMappingsAck {}
// SyncMappingsResponse is a batch of mappings sent by management.
// Identical semantics to GetMappingUpdateResponse.
message SyncMappingsResponse {
repeated ProxyMapping mapping = 1;
// initial_sync_complete is set on the last message of the initial snapshot.
bool initial_sync_complete = 2;
}