Close search's active-tenant write-routing gap with a polled allowlist

search/src/consumer.rs's write-routing (built last pass) had no active-
tenant check at all: IndexRegistry.resolve() would open-or-create an
index directory for any syntactically-valid tenant_id, active or not --
unlike ClickHouse's chwriter.Registry (an active-tenants-only snapshot
built at enterprise-ingest startup) or the read side (gated by
searchclient.TenantChecker, a direct rbacstore query). search is AGPL
core with no Postgres access and no enterprise/ import allowed, so it
needed a network boundary instead -- the same shape ingest's
TenantResolver already uses against enterprise-auth, just Rust calling
Go instead of Go calling Go.

New GET /internal/active-tenants endpoint on enterprise-auth
(rbacstore.ListActiveTenantIDs + authhandler.handleActiveTenants),
gated on a RoleService Bearer credential -- server-to-server auth, the
same shape alerting presents to api, minted via the already-generic
enterprise-auth -mint-service-token search. search/src/tenants.rs's
ActiveTenantTracker polls it every 60s, blocking startup on the first
fetch succeeding (fail-closed cold start -- a control-plane outage at
boot must not silently accept every tenant_id) and keeping the last-
known-good set on any later refresh failure (a transient blip shouldn't
stop every tenant's indexing, only prevent the allowlist from growing/
shrinking until connectivity resumes). consumer.rs refuses any tagged
record whose tenant isn't in the polled set, before ever calling
resolve() -- IndexRegistry itself stays policy-free, matching the same
mechanism/policy split clickhousewriter.Writer vs. chwriter.Registry
already draws on the ClickHouse side.

Off unless ENTERPRISE_AUTH_URL/ENTERPRISE_AUTH_SERVICE_TOKEN are both
set (search/src/config.rs rejects exactly one being set) -- every
existing deployment is unaffected.

Verified with real HTTP round trips in this environment: tenants.rs's
tests exercise real reqwest requests (actual Authorization: Bearer
header, actual JSON parsing) against a hand-rolled dependency-free TCP
test server, including both fail-closed paths (rejected first fetch,
unreachable server). authhandler's new tests cover the credential-kind
distinction this endpoint exists to enforce -- a real human session,
even for a genuine Owner, must not satisfy a check meant for a service
identity.

One asymmetry remains, disclosed rather than fixed: chwriter.Registry's
snapshot still never refreshes (stale until enterprise-ingest restarts),
while ActiveTenantTracker's 60s poll gives Tantivy a materially tighter
staleness window. Neither is a live per-write check -- that would mean
a database/HTTP round trip per record, a throughput cost neither
implementation accepts -- so both have some staleness window by design;
the gap between the two windows is what's disclosed, not a claim either
is fully live.
This commit is contained in:
2026-08-14 23:47:25 -07:00
parent 5a845f06ee
commit 088677643f
19 changed files with 1139 additions and 141 deletions
+58 -2
View File
@@ -13,6 +13,9 @@
// api/authz) and a different credential type (an ingest bearer token
// checked against rbacstore's ingest_credentials table, not a
// session.Manager JWT) -- see ingestCredentialValidator's doc comment.
//
// GET /internal/active-tenants is a third sibling, for `search`
// (core/AGPL, Rust) -- see tenantLister's doc comment.
package authhandler
import (
@@ -50,21 +53,36 @@ type ingestCredentialValidator interface {
ValidateIngestCredential(ctx context.Context, token string) (tenantID string, err error)
}
// tenantLister is the narrow interface GET /internal/active-tenants
// needs -- *rbacstore.Store is the production implementation (the same
// concrete type ingestCredentials above already wires in, just a
// second narrow interface it happens to also satisfy). Backs
// search/src/tenants.rs's ActiveTenantTracker: search is AGPL core with
// no Postgres access and no enterprise/ import allowed, so its write-
// routing needed a network boundary to learn which tenants are active,
// the same shape ingest/internal/grpcserver.TenantResolver already uses
// against this exact service (see that package's doc comment).
type tenantLister interface {
ListActiveTenantIDs(ctx context.Context) ([]string, error)
}
type Handler struct {
logger *slog.Logger
manager *session.Manager
features Features
ingestCredentials ingestCredentialValidator
tenants tenantLister
}
func New(logger *slog.Logger, manager *session.Manager, features Features, ingestCredentials ingestCredentialValidator) *Handler {
return &Handler{logger: logger, manager: manager, features: features, ingestCredentials: ingestCredentials}
func New(logger *slog.Logger, manager *session.Manager, features Features, ingestCredentials ingestCredentialValidator, tenants tenantLister) *Handler {
return &Handler{logger: logger, manager: manager, features: features, ingestCredentials: ingestCredentials, tenants: tenants}
}
func (h *Handler) RegisterRoutes(mux *http.ServeMux) {
mux.HandleFunc("POST /internal/authorize", h.handleAuthorize)
mux.HandleFunc("GET /auth/features", h.handleFeatures)
mux.HandleFunc("POST /internal/authorize-ingest", h.handleAuthorizeIngest)
mux.HandleFunc("GET /internal/active-tenants", h.handleActiveTenants)
}
type featuresResponse struct {
@@ -146,6 +164,44 @@ func (h *Handler) handleAuthorizeIngest(w http.ResponseWriter, r *http.Request)
_ = json.NewEncoder(w).Encode(authorizeIngestResponse{TenantID: tenantID})
}
type activeTenantsResponse struct {
TenantIDs []string `json:"tenant_ids"`
}
// handleActiveTenants is search/src/tenants.rs's ActiveTenantTracker's
// server side -- polled periodically, not per-write, to build a local
// allowlist for its write-routing gate (see that module's doc comment).
// Requires a RoleService Bearer credential (session.Manager-issued,
// Role == "service"), not a human session -- server-to-server, the same
// authentication shape /alerting presents to /api, minted via
// `enterprise-auth -mint-service-token search` (the flag is already
// generic over caller name; no change needed there for a new caller).
// Deliberately does NOT accept the tenant-scoped credential a human
// session or an ingest credential would carry: this endpoint answers
// "which tenants exist," a question no single tenant's identity should
// be able to ask on its own.
func (h *Handler) handleActiveTenants(w http.ResponseWriter, r *http.Request) {
token := bearerToken(r.Header.Get("Authorization"))
if token == "" {
http.Error(w, "no credentials presented", http.StatusUnauthorized)
return
}
claims, err := h.manager.Validate(token)
if err != nil || claims.Role != "service" {
http.Error(w, "invalid or expired credentials", http.StatusUnauthorized)
return
}
ids, err := h.tenants.ListActiveTenantIDs(r.Context())
if err != nil {
h.logger.Error("listing active tenant ids", "error", err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
w.Header().Set("Content-Type", "application/json")
_ = json.NewEncoder(w).Encode(activeTenantsResponse{TenantIDs: ids})
}
func bearerToken(header string) string {
const prefix = "Bearer "
if !strings.HasPrefix(header, prefix) {