---
id: review-001-gateway-stream-errors
name: Gateway error-fidelity — SSE terminality, mappings, Retry-After, deadline (GW-03..GW-07, GW-12..GW-14)
status: completed
depends_on: []
scope: narrow
risk: low
impact: component
level: implementation
tags: [gateway, review-001]
---

## Description

Review 001 gateway error-fidelity findings
(`src/gateway/routes.rs`, `error.rs`, `dispatch.rs`) — the response a
caller sees must match the documented contract and match across
transports:

- **GW-04**: SSE error events are not terminal — the stream continues
  after `Err` (`routes.rs:299-313`) while the wire dispatcher treats
  `Err` as terminal; two transports disagree about stream semantics
  (http-server.md:219-223 documents terminal). Emit the error frame and
  end the stream (`take_while`).
- **GW-03**: `INVALID_OPERATION_TYPE` → 500 on `/call`//`/batch` (client
  fault reported as server fault; pollutes alerting) but 400 on
  `/publish`. Map consistently (400 or 422) and document in
  http-server.md.
- **GW-05**: the 30 s `DEFAULT_TIMEOUT` deadline is recorded
  (`dispatch.rs:34,165`) but never enforced — a hung handler holds the
  request open indefinitely. Either enforce (`tokio::time::timeout`
  around Once-op invokes) or remove the dead metadata. Prefer enforcing;
  it is the documented contract.
- **GW-07**: `Retry-After` machinery exists in
  `call_error_to_http_response` (`error.rs:61-75`) but the main gateway
  error path builds responses by hand (`routes.rs:319-333`) — retryable
  `HTTP_429`/`HTTP_503` reach callers with no `Retry-After` despite the
  documented mapping. Route gateway error responses through the shared
  mapper.
- **GW-12**: ACL denial on `/subscribe` surfaces as HTTP 200 +
  `event:error` while `/call` returns 401/403. ADR-049 makes
  200-on-stream defensible, but the doc must call out the asymmetry —
  coordinate with review-001-output-projection (PRJ-05) which documents
  it.
- **GW-13**: no SSE keep-alive/heartbeat (`routes.rs:172`) — quiet-but-
  alive streams (the normal state for subscriptions) die at LB/proxy
  idle timeouts. Add keep-alive + `retry:` field.
- **GW-14**: stale module doc (`routes.rs:1-9` claims `/publish` is "a
  separate module"; AGENTS.md §7's "5 gateway endpoints" framing is
  similarly stale — ADR-068 made it 6).

## Acceptance Criteria

- [x] SSE stream ends after an error event (test); keep-alive present
- [x] `INVALID_OPERATION_TYPE` maps to the same status on `/call`, `/batch`, `/publish`; http-server.md table updated
- [x] Once-op invokes enforce the 30 s deadline (test with a hung handler)
- [x] Retryable errors carry `Retry-After` on all live error paths (test)
- [x] GW-14 docs fixed (module doc; AGENTS.md gateway-endpoint count if touched)
- [x] `cargo test` and `cargo clippy --all-targets -- -D warnings` pass

## References

- docs/reviews/001-initial-implementation-review.md (Part C, GW-03..GW-07, GW-12..GW-14)
- docs/architecture/decisions/049-streaming-handler-for-subscriptions.md
- docs/architecture/decisions/023-operation-error-schemas.md

## Notes

> Agent fills during implementation. Independent of
> review-001-gateway-publish-semantics (both touch routes.rs —
> sequence or coordinate to avoid churn).

Built on d7ee302 (GW-01/GW-06/HY-13 publish semantics + the error.rs
`INVALID_OPERATION_TYPE` mapper) — not re-implemented or reverted.
GW-03 needed no new mapping: the publish-side mapper from d7ee302
(`422` with identity / `401` without) already covers `/call` and
`/batch` because both route their error envelopes through the shared
mapper; documented in http-server.md's error table.

## Summary

- **GW-04**: `subscribe_stream_from_envelope_stream` reframed with a
  `scan`-based projection — the first `Err` envelope's `event:error`
  frame is emitted, then the stream terminates (the scan flags done and
  subsequent polls return `None`). New test
  `subscribe_stream_is_terminal_after_an_error_event` proves a
  post-error envelope never reaches the wire. Matches the wire
  dispatcher's `call.error`-is-terminal semantics.
- **GW-03**: no code change needed beyond d7ee302's mapper — verified
  consistent `422`/`401` across `/call`/`/batch`/`/publish` via the
  existing `error_streaming_handler`-style tests plus the mapper's unit
  tests; http-server.md error table now documents
  `INVALID_OPERATION_TYPE` explicitly (422 with token / 401 without).
- **GW-05**: enforced, not removed. `GatewayDispatch::invoke` wraps the
  registry invoke in `tokio::time::timeout(DEFAULT_TIMEOUT)`; a hung
  handler yields a `TIMEOUT` error envelope (`504`,
  `retryable: true`) instead of holding the request open. Streaming and
  sink invokes stay unbounded (ADR-021 sets `deadline: None` for
  subscriptions; a `/publish` body is upload-bounded). Tests: hung
  handler (120 s sleep) returns well under 60 s with `TIMEOUT`; fast
  handler unaffected. The `deadline:` `OperationContext` metadata stays
  (it feeds registry-side bookkeeping).
- **GW-07**: all live gateway error paths now route through the shared
  mapper — `envelope_to_response` and `forbidden_response` call the new
  `call_error_to_http_response_with_identity` (added in error.rs,
  identity-aware variant of the existing builder that keeps the 401/403
  and 401/422 splits). Test: a retryable `HTTP_503` handler error on
  `/call` carries `Retry-After: 30`.
- **GW-12**: behavior unchanged (200-on-stream per ADR-049); the
  asymmetry is now documented in http-server.md's streaming section
  ("Status-code asymmetry on /subscribe").
- **GW-13**: `subscribe_handler` returns
  `Sse::new(stream).keep_alive(KeepAlive…15 s)` with a `retry:` field
  (15000 ms) on every stream event and on the keep-alive comment frame
  itself (axum's event-based keep-alive re-sends the full event, so
  quiet streams still deliver a `retry` hint). Interval is the
  `SSE_KEEP_ALIVE_INTERVAL` constant (15 s), documented in http-server.md.
- **GW-14**: routes.rs module doc rewritten (6 endpoints, `/publish` in
  this module, stale "separate module" claim removed). AGENTS.md §7's
  "5 gateway endpoints" framing was NOT edited — AGENTS.md is a parallel
  read for other agents; the operative count is documented in
  http-server.md and ADR-068 (AGENTS.md alignment can ride a docs-only
  commit).

Verification: cargo test (243 passed, incl. 9 new); cargo clippy
--all-targets -- -D warnings clean; cargo fmt --check clean.