[management,proxy] Agent network: per-account LLM gateway (policy, metering, multi-provider) (#6555)

* [agent-network] Shared proto, OpenAPI schema, and generated types

* [agent-network] Management: store, manager, synthesizer, policy engine, provider catalog, HTTP/gRPC API

Adds the account-scoped agent-network module: provider/policy/budget CRUD and
store, the reverse-proxy service synthesizer, policy selection + limit
enforcement, the provider catalog (incl. Vertex AI and AWS Bedrock entries),
and the management HTTP + proxy gRPC surfaces.

* [management] Fix agent-network proxy-peer fan-out on affected-peer recompute

The affected-peers resolver loaded only persisted reverse-proxy services, but
agent-network services are synthesized on demand and never persisted. As a
result the embedded proxy peer was never folded into the affected set when a
client's group changed, so the proxy received no network-map update for a newly
authorised client and rejected its handshake until a full resync (restart).

loadProxyServices now merges the synthesized agent-network services (injected
via a registration hook to avoid an import cycle), so proxy peers learn newly
authorised clients immediately.

* [proxy] Reverse-proxy middleware framework, chain, and request plumbing

The per-target middleware chain (slots, dispatcher, mutation gate, metadata
merger), body capture, access-log terminal sink, and the proxy wiring that
builds + runs chains for synthesized agent-network services.

* [proxy] LLM parsers, pricing, and builtin middlewares (OpenAI, Anthropic, Vertex AI, AWS Bedrock)

Request/response parsers and SSE/event-stream metering, the embedded pricing
table, and the builtin middleware set: request parser, router, policy
limit-check/record, cost meter, guardrail, identity inject, response parser.
Includes the path-routed providers — Google Vertex AI (keyfile:: service-account
OAuth minting) and AWS Bedrock (bearer auth, invoke/converse/streaming, optional
/bedrock prefix) — plus the Models allowlist and unmeterable-publisher deny.

* [proxy] IPv6 in-place apply and TCP accept-loop hardening on netstack listeners

* [agent-network] End-to-end test suite, module docs, and deployment preset

* [agent-network] Fix codespell typos and exclude false positives

- labelgen word pool: vermillion -> vermilion, racoon -> raccoon.
- codespell ignore list: add flate (Go compress/flate package), recordin
  (a test-local identifier), and unparseable (a valid alternative spelling used
  consistently across identifiers + a metadata-value constant).

* [management] Set LastSeen on injected proxy peer in realstack test (MySQL strict-mode)

The injected embedded proxy peer had a PeerStatus with a zero LastSeen, which
serializes to '0000-00-00' and is rejected by MySQL in strict mode (SQLite
tolerates it). Set LastSeen to a valid time so SaveAccount succeeds on both
engines.

* [agent-network] Remove e2e shell-script suite from this branch

The end-to-end shell scripts under scripts/e2e/ are maintained in a separate
testing suite and are not part of this change set.

* [agent-network] Polish module docs: remove internal review scaffolding, fix links, verify diagrams

Strip PR-review framing, commit references, absolute paths, and stale internal
references from the agent-network module docs; fix broken relative links; verify
all diagrams against the current architecture. Remove the internal AI-reviewer
prompt file.

* [management] Refine session expiration handling to support 3-state encoding for SSO deadlines

* [agent-network] Relocate agentnetwork package to internals/modules

Move management/server/agentnetwork (and its catalog/, labelgen/, types/
subpackages) to management/internals/modules/agentnetwork, alongside the
reverse-proxy module, and rewrite all importers. Pure relocation: package names,
the synthesizer + affectedpeers registration hook, and store access (shared
store.Store) are unchanged, so no import cycle is introduced (affectedpeers
still depends only on the agentnetwork/types leaf).

* [agent-network] Co-locate HTTP handlers in the module (RegisterEndpoints)

Move the agent-network HTTP handlers from server/http/handlers/agentnetwork into
the module at internals/modules/agentnetwork/handlers (package handlers) and
rename the entrypoint AddEndpoints -> RegisterEndpoints, matching the
reverse-proxy module convention. Wiring in http/handler.go updated accordingly.
This commit is contained in:
Maycon Santos
2026-06-27 13:41:00 +02:00
committed by GitHub
parent 615631567a
commit b416063bcc
187 changed files with 36835 additions and 660 deletions
@@ -43,6 +43,16 @@ type ProxyServiceClient interface {
// issue a session cookie without redirecting through the OIDC flow.
// Mirrors ValidateSession's response shape.
ValidateTunnelPeer(ctx context.Context, in *ValidateTunnelPeerRequest, opts ...grpc.CallOption) (*ValidateTunnelPeerResponse, error)
// CheckLLMPolicyLimits is the pre-flight RPC the proxy calls before each
// LLM request. Management runs the per-policy headroom selection across
// every policy authorising the caller's user / groups for the resolved
// provider and returns the chosen attribution policy + group, or a deny
// when no applicable policy has headroom > 0.
CheckLLMPolicyLimits(ctx context.Context, in *CheckLLMPolicyLimitsRequest, opts ...grpc.CallOption) (*CheckLLMPolicyLimitsResponse, error)
// RecordLLMUsage is the post-flight RPC the proxy calls after the upstream
// returns. Increments the per-(dimension, window) counters for the
// attribution policy chosen by CheckLLMPolicyLimits.
RecordLLMUsage(ctx context.Context, in *RecordLLMUsageRequest, opts ...grpc.CallOption) (*RecordLLMUsageResponse, error)
}
type proxyServiceClient struct {
@@ -179,6 +189,24 @@ func (c *proxyServiceClient) ValidateTunnelPeer(ctx context.Context, in *Validat
return out, nil
}
func (c *proxyServiceClient) CheckLLMPolicyLimits(ctx context.Context, in *CheckLLMPolicyLimitsRequest, opts ...grpc.CallOption) (*CheckLLMPolicyLimitsResponse, error) {
out := new(CheckLLMPolicyLimitsResponse)
err := c.cc.Invoke(ctx, "/management.ProxyService/CheckLLMPolicyLimits", in, out, opts...)
if err != nil {
return nil, err
}
return out, nil
}
func (c *proxyServiceClient) RecordLLMUsage(ctx context.Context, in *RecordLLMUsageRequest, opts ...grpc.CallOption) (*RecordLLMUsageResponse, error) {
out := new(RecordLLMUsageResponse)
err := c.cc.Invoke(ctx, "/management.ProxyService/RecordLLMUsage", in, out, opts...)
if err != nil {
return nil, err
}
return out, nil
}
// ProxyServiceServer is the server API for ProxyService service.
// All implementations must embed UnimplementedProxyServiceServer
// for forward compatibility
@@ -208,6 +236,16 @@ type ProxyServiceServer interface {
// issue a session cookie without redirecting through the OIDC flow.
// Mirrors ValidateSession's response shape.
ValidateTunnelPeer(context.Context, *ValidateTunnelPeerRequest) (*ValidateTunnelPeerResponse, error)
// CheckLLMPolicyLimits is the pre-flight RPC the proxy calls before each
// LLM request. Management runs the per-policy headroom selection across
// every policy authorising the caller's user / groups for the resolved
// provider and returns the chosen attribution policy + group, or a deny
// when no applicable policy has headroom > 0.
CheckLLMPolicyLimits(context.Context, *CheckLLMPolicyLimitsRequest) (*CheckLLMPolicyLimitsResponse, error)
// RecordLLMUsage is the post-flight RPC the proxy calls after the upstream
// returns. Increments the per-(dimension, window) counters for the
// attribution policy chosen by CheckLLMPolicyLimits.
RecordLLMUsage(context.Context, *RecordLLMUsageRequest) (*RecordLLMUsageResponse, error)
mustEmbedUnimplementedProxyServiceServer()
}
@@ -242,6 +280,12 @@ func (UnimplementedProxyServiceServer) ValidateSession(context.Context, *Validat
func (UnimplementedProxyServiceServer) ValidateTunnelPeer(context.Context, *ValidateTunnelPeerRequest) (*ValidateTunnelPeerResponse, error) {
return nil, status.Errorf(codes.Unimplemented, "method ValidateTunnelPeer not implemented")
}
func (UnimplementedProxyServiceServer) CheckLLMPolicyLimits(context.Context, *CheckLLMPolicyLimitsRequest) (*CheckLLMPolicyLimitsResponse, error) {
return nil, status.Errorf(codes.Unimplemented, "method CheckLLMPolicyLimits not implemented")
}
func (UnimplementedProxyServiceServer) RecordLLMUsage(context.Context, *RecordLLMUsageRequest) (*RecordLLMUsageResponse, error) {
return nil, status.Errorf(codes.Unimplemented, "method RecordLLMUsage not implemented")
}
func (UnimplementedProxyServiceServer) mustEmbedUnimplementedProxyServiceServer() {}
// UnsafeProxyServiceServer may be embedded to opt out of forward compatibility for this service.
@@ -428,6 +472,42 @@ func _ProxyService_ValidateTunnelPeer_Handler(srv interface{}, ctx context.Conte
return interceptor(ctx, in, info, handler)
}
func _ProxyService_CheckLLMPolicyLimits_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
in := new(CheckLLMPolicyLimitsRequest)
if err := dec(in); err != nil {
return nil, err
}
if interceptor == nil {
return srv.(ProxyServiceServer).CheckLLMPolicyLimits(ctx, in)
}
info := &grpc.UnaryServerInfo{
Server: srv,
FullMethod: "/management.ProxyService/CheckLLMPolicyLimits",
}
handler := func(ctx context.Context, req interface{}) (interface{}, error) {
return srv.(ProxyServiceServer).CheckLLMPolicyLimits(ctx, req.(*CheckLLMPolicyLimitsRequest))
}
return interceptor(ctx, in, info, handler)
}
func _ProxyService_RecordLLMUsage_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
in := new(RecordLLMUsageRequest)
if err := dec(in); err != nil {
return nil, err
}
if interceptor == nil {
return srv.(ProxyServiceServer).RecordLLMUsage(ctx, in)
}
info := &grpc.UnaryServerInfo{
Server: srv,
FullMethod: "/management.ProxyService/RecordLLMUsage",
}
handler := func(ctx context.Context, req interface{}) (interface{}, error) {
return srv.(ProxyServiceServer).RecordLLMUsage(ctx, req.(*RecordLLMUsageRequest))
}
return interceptor(ctx, in, info, handler)
}
// ProxyService_ServiceDesc is the grpc.ServiceDesc for ProxyService service.
// It's only intended for direct use with grpc.RegisterService,
// and not to be introspected or modified (even as a copy)
@@ -463,6 +543,14 @@ var ProxyService_ServiceDesc = grpc.ServiceDesc{
MethodName: "ValidateTunnelPeer",
Handler: _ProxyService_ValidateTunnelPeer_Handler,
},
{
MethodName: "CheckLLMPolicyLimits",
Handler: _ProxyService_CheckLLMPolicyLimits_Handler,
},
{
MethodName: "RecordLLMUsage",
Handler: _ProxyService_RecordLLMUsage_Handler,
},
},
Streams: []grpc.StreamDesc{
{