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:
+21
-23
@@ -22,29 +22,27 @@ use crate::index::SearchIndex;
|
||||
/// path every Phase 0-3 deployment, and every untagged record, still
|
||||
/// uses.
|
||||
///
|
||||
/// **Known residual gap, disclosed rather than silently accepted**:
|
||||
/// unlike the read side (gated by `enterprise/internal/searchclient`'s
|
||||
/// `TenantChecker`, which refuses to even issue a search for a tenant
|
||||
/// that isn't `active` in `rbacstore`) and unlike ClickHouse's write
|
||||
/// side (`enterprise/internal/chwriter.Registry`, built once at startup
|
||||
/// from `rbacstore.ListProvisionedDataSources` -- `active` tenants
|
||||
/// only, so an unrecognized `tenant_id` has no writer and the whole
|
||||
/// batch is refused), this registry's `resolve` has no equivalent gate
|
||||
/// on the write path: `consumer.rs` calls it directly, with no Postgres
|
||||
/// access to check tenant status against, the same reason this
|
||||
/// module's doc comment used to give for the old read-side gap
|
||||
/// `TenantChecker` was built to close. A syntactically-valid `tenant_id`
|
||||
/// on an ingest credential that's still valid but should have been
|
||||
/// revoked (deprovisioning does not yet revoke `ingest_credentials`
|
||||
/// rows -- see `/CLAUDE.md`'s Phase 4 non-goals) can therefore cause an
|
||||
/// index directory to be silently created here for a tenant that isn't
|
||||
/// really active. The blast radius is narrow -- an orphan, isolated,
|
||||
/// empty-except-for-that-tenant's-own-traffic index directory, not
|
||||
/// cross-tenant data exposure, and only reachable with a real signed
|
||||
/// ingest credential, not by an arbitrary caller -- but it is real, not
|
||||
/// hypothetical. Closing it fully would mean giving `search` (AGPL
|
||||
/// core, no `enterprise/` import allowed) some way to learn which
|
||||
/// tenants are actually active; not designed yet.
|
||||
/// `resolve` itself still has no active-tenant gate of its own -- it
|
||||
/// will happily open-or-create an index for any syntactically-valid
|
||||
/// `tenant_id`, active or not. That's deliberate: this struct's job is
|
||||
/// managing index lifecycles, not policy, the same separation
|
||||
/// `clickhousewriter.Writer` (mechanism) vs. `chwriter.Registry`
|
||||
/// (policy: which tenants get a writer at all) draws on the ClickHouse
|
||||
/// side. The gate lives one layer up, at each caller:
|
||||
/// `enterprise/internal/searchclient.TenantChecker` for the read side
|
||||
/// (refuses to even issue a search for a tenant that isn't `active` in
|
||||
/// `rbacstore`, via a direct Postgres-backed query since that code runs
|
||||
/// in `enterprise/`), and `consumer.rs`'s `tenants::ActiveTenantTracker`
|
||||
/// for the write side (a polled allowlist fetched from a new
|
||||
/// `enterprise-auth` endpoint over HTTP -- `search` is AGPL core with no
|
||||
/// Postgres access and no `enterprise/` import allowed, so it needed a
|
||||
/// network boundary instead of an import one, the same shape
|
||||
/// `ingest/internal/grpcserver.TenantResolver` already uses against the
|
||||
/// same service). Both gates are optional at this layer -- `resolve`
|
||||
/// itself works identically whether or not either caller happens to
|
||||
/// gate it -- so a future caller that forgets to gate would silently
|
||||
/// reopen this exact class of gap; see `consumer.rs`'s call site for
|
||||
/// the write side's enforcement.
|
||||
pub struct IndexRegistry {
|
||||
default_index: Arc<SearchIndex>,
|
||||
tenants_root: PathBuf,
|
||||
|
||||
Reference in New Issue
Block a user