fix(security): partition the response cache by credential and principal (#3862)

## Why

On a proxy shared by several callers, the semantic response cache was
keyed only by request content and upstream (#3349). A second caller who
sent the identical request with a different provider key was served the
completion generated for the first caller, under the first caller's key,
and never reached the upstream. That discloses completions across
tenants (finding 01-F15).

## What changes

- `compute_cache_partition()` in
`headroom/proxy/semantic_cache_key_policy.py`: an HMAC-SHA256 under a
per-process random key over the caller's credential and account-selector
headers plus the authenticated principal. Headers counted: every header
the shared credential-header rule
(`internal_header_policy.is_credential_header`) matches, which covers
`authorization`, `proxy-authorization`, `cookie` and `x-api-key`, and
any header whose name ends in `api-key`, `key`, `token`, `secret`,
`account`, `account-id`, `organization` or `project` (so `x-api-key`,
`api-key`, `x-goog-api-key`, `chatgpt-account-id`,
`openai-organization`, `openai-project` and so on). Matching by name
shape keeps new providers partitioned by default. Cookie is used as a
whole value; it is only ever HMAC input. `x-headroom-*` headers
(including the proxy's own token, which the security gate has already
removed) and idempotency keys are excluded. A request with no credential
and no principal gets the shared `anon` partition.
- The partition is an HMAC, not a plain hash, so it cannot be used to
confirm a guessed credential.
- `identity.resolve_authenticated_principal()` returns a principal only
when an identity resolver is installed (`set_identity_resolver`). The
OSS default identity is the same for every network caller and can be
chosen by header on loopback, so it is not treated as a principal.
- If an installed identity resolver raises or returns no principal, the
request fails closed (`resolve_authenticated_principal()` raises
`UnresolvedPrincipalError` for an empty result):
`compute_request_cache_partition()` logs a warning and returns `None`,
and both handlers skip the response cache for that request (no lookup,
no store). It never falls back to the credential-only partition, which
tenants on one operator key would share. With no resolver installed, the
credential-only partition is unchanged.
- Both handlers compute the partition only when the response cache can
be used (cache enabled and non-streaming), so streaming and
cache-disabled requests do not call the identity resolver.
- `partition` is a required keyword on `SemanticCache.get`, `set` and
`_compute_key`, and on `compute_semantic_cache_key`. A future call site
that forgets it fails immediately instead of silently sharing.
- The `/v1/messages` and `/v1/chat/completions` handlers compute the
partition once, next to the existing cache-key snapshot, and reuse it at
store time so lookup and store agree.

Other caches in the request path are not changed. The OpenAI Responses
unit cache and the CCR retrieval store are content-addressed, so a hit
requires already holding the content. CCR entries are not bound to a
caller; that is a separate issue. The `/v1/compress` pipeline cache
holds no user content, and the SDK `SemanticCache` is in-process for one
caller.

## Behaviour change

The proxy response cache no longer shares entries between callers that
present different provider credentials or, with an identity resolver
installed, different principals.

- One developer with one key: same hit rate.
- Deployments where callers send no provider credential and the proxy
supplies the operator's key: same hit rate; they share the `anon`
partition.
- Shared proxy with per-caller keys: entries are no longer shared
between keys. That is the fix.
- Compatible upstream gateways that authenticate by `Cookie` or
`Proxy-Authorization`: entries are no longer shared between sessions.
- Callers whose only credential is a proxy-token `Authorization` header,
or deployments using oauth2, share one partition unless an identity
resolver is installed, because core cannot tell them apart.

## Tests

`tests/test_proxy_response_cache_partition.py`:
- The partition function: differs per credential and is stable for one
credential; every credential and account header shape partitions; other
headers do not fragment the cache; no credential and no principal gives
the shared partition; it is keyed, not a plain hash; a principal
separates callers sharing one credential.
- `SemanticCache` refuses a call without a partition and keeps
partitions apart.
- End to end through the real Anthropic and OpenAI handlers: a second
caller is never served the first caller's cached response, whether the
credential is `x-api-key`/`Authorization`, `Cookie` or
`Proxy-Authorization`. The `Cookie` and `Proxy-Authorization` cases fail
without the latest change. With an identity resolver installed, callers
on one operator key are partitioned. These end-to-end tests fail on
`main`.
- A resolver that raises, or returns an empty principal, yields no
partition, and end to end (both handlers) callers on one operator key
never share a cached response and nothing is stored. These fail without
the fail-closed change. With no resolver installed, the partition is the
credential-only one.
- With the response cache disabled, neither handler calls the identity
resolver.

Existing cache tests now pass a fixed partition. The buffered-stream CCR
test derives the same partition the handler computes.

Manual run: two callers send the identical `/v1/messages` request with
different `x-api-key` values through a proxy in front of a capture
upstream. On `main` the second caller gets the first caller's answer and
only the first key reaches the upstream. On this branch each caller gets
its own answer, and a repeat by the first caller is still a cache hit.

Not tested: multi-worker deployments (the cache and its key are per
process, as before).

## Merge order / conflicts

Part of a set. Suggested order: #3852 → #3891 → #3849 → #3860 → #3862 →
#3863.

Textual conflict: #3860 adds an import next to the one this PR adds in
`handlers/anthropic.py` and `handlers/openai.py`. Whichever merges
second keeps both imports.

Revised 2026-10-04 after review: forwarded `Cookie` and
`Proxy-Authorization` headers are now partition input (via the shared
credential-header rule), so cookie-authenticated callers no longer share
the anonymous partition.

🤖 Generated with [Claude Code](https://claude.com/claude-code)


<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/headroomlabs-ai/headroom/pull/3862"
target="_blank"><picture><source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-devin-review-dark.svg?v=4"><img
src="https://static.devin.ai/assets/gh-devin-review-light.svg?v=4"
alt="Devin Review"></picture></a>
<!-- devin-review-badge-end -->

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: JD Davis <mxjerrett@gmail.com>
This commit is contained in:
Tejas Chopra
2026-10-05 15:19:08 -05:00
committed by GitHub
co-authored by Claude Opus 5.5 JD Davis
parent 6151ed1f52
commit ee6cfcccf6
10 changed files with 612 additions and 29 deletions
+16 -2
View File
@@ -66,6 +66,7 @@ from headroom.proxy.nonstream_sse_policy import should_recover_sse_reply
from headroom.proxy.outcome import RequestOutcome
from headroom.proxy.output_shaper import shaper_enabled_for, steering_allowed_for
from headroom.proxy.rate_limit_identity import rate_limit_identity
from headroom.proxy.semantic_cache_key_policy import compute_request_cache_partition
from headroom.proxy.tenant_key import resolve_tenant_key, set_request_tenant_key
from headroom.proxy.thinking_tokens import ThinkingTokens, extract_thinking_tokens
from headroom.utils import format_exception_message
@@ -1430,10 +1431,21 @@ class AnthropicHandlerMixin:
# unreachable entries. Reuse this raw snapshot verbatim at cache.set
# (the same reason cache_key_fields is snapshotted here, #327).
cache_lookup_messages = messages
# Response-cache partition: a cached response is only ever replayed to a
# caller presenting the same provider credentials and principal (01-F15).
# Snapshotted with the key fields so lookup and store agree. None means
# the principal could not be established: skip the cache entirely.
# Only resolved when the cache can be used, so streaming and
# cache-disabled requests never pay for identity resolution.
cache_partition = (
compute_request_cache_partition(request) if self.cache and not stream else None
)
# Check cache (non-streaming only)
cache_hit = False
if self.cache and not stream:
cached = await self.cache.get(messages, model, **cache_key_fields)
if self.cache and not stream and cache_partition is not None:
cached = await self.cache.get(
messages, model, partition=cache_partition, **cache_key_fields
)
if cached:
cache_hit = True
self.pipeline_extensions.emit(
@@ -4835,6 +4847,7 @@ class AnthropicHandlerMixin:
if (
self.cache
and not stream
and cache_partition is not None
and response.status_code == 200
and resp_json is not None
):
@@ -4844,6 +4857,7 @@ class AnthropicHandlerMixin:
response.content,
dict(response.headers),
tokens_saved=tokens_saved,
partition=cache_partition,
**cache_key_fields,
)
+21 -3
View File
@@ -37,6 +37,7 @@ from headroom.proxy.identity import resolve_memory_identity
from headroom.proxy.loopback_guard import is_loopback_host
from headroom.proxy.modes import is_cache_mode
from headroom.proxy.rate_limit_identity import rate_limit_identity
from headroom.proxy.semantic_cache_key_policy import compute_request_cache_partition
from headroom.proxy.stage_timer import StageTimer, emit_stage_timings_log
from headroom.proxy.upstream_guard import is_safe_upstream_url
from headroom.proxy.ws_headers import WS_HOP_BY_HOP_HEADERS
@@ -3915,9 +3916,20 @@ class OpenAIHandlerMixin:
# captured — keep this snapshot after image compression, or a reorder
# silently reintroduces the drift.
cache_lookup_messages = messages
# Response-cache partition: a cached response is only ever replayed to a
# caller presenting the same provider credentials and principal (01-F15).
# Snapshotted with the key fields so lookup and store agree. None means
# the principal could not be established: skip the cache entirely.
# Only resolved when the cache can be used, so streaming and
# cache-disabled requests never pay for identity resolution.
cache_partition = (
compute_request_cache_partition(request) if self.cache and not stream else None
)
# Check cache
if self.cache and not stream:
cached = await self.cache.get(messages, model, **cache_key_fields)
if self.cache and not stream and cache_partition is not None:
cached = await self.cache.get(
messages, model, partition=cache_partition, **cache_key_fields
)
if cached:
self.pipeline_extensions.emit(
PipelineStage.INPUT_CACHED,
@@ -5802,13 +5814,19 @@ class OpenAIHandlerMixin:
# site, which let a response built for a stream:true request
# answer a later non-streaming caller (#3019). Stating the
# invariant keeps that from being reintroduced silently.
if self.cache and not stream and response.status_code == 200:
if (
self.cache
and not stream
and cache_partition is not None
and response.status_code == 200
):
await self.cache.set(
cache_lookup_messages,
model,
response.content,
dict(response.headers),
tokens_saved,
partition=cache_partition,
**cache_key_fields,
)
+27
View File
@@ -30,12 +30,39 @@ class IdentityResolver(Protocol):
_resolver: IdentityResolver | None = None
class UnresolvedPrincipalError(LookupError):
"""An installed identity resolver could not establish a principal."""
def set_identity_resolver(resolver: IdentityResolver | None) -> None:
"""Install (or clear) a custom identity resolver — the enterprise hook."""
global _resolver
_resolver = resolver
def resolve_authenticated_principal(request: Any) -> str | None:
"""Return the authenticated principal for ``request``, or ``None``.
Only a registered custom resolver (see :func:`set_identity_resolver`)
establishes a real per-caller identity. The OSS default binds every
network caller to the same proxy-token identity and lets loopback callers
*choose* a partition by header, so neither is an authenticated principal
and both return ``None`` here. Callers that must isolate data between
principals (e.g. the response cache) combine this with the caller's own
provider credential rather than trusting the default identity.
An installed resolver that raises, or returns no principal, means the
principal is unknown rather than absent: this raises so the caller fails
closed instead of silently sharing data across principals.
"""
if _resolver is None:
return None
principal = _resolver(request, default="")
if not principal:
raise UnresolvedPrincipalError("identity resolver returned no principal")
return principal
def _default_os_user() -> str:
return os.environ.get("USER", os.environ.get("USERNAME", "default"))
+17 -6
View File
@@ -61,7 +61,9 @@ class SemanticCache:
self._cache: OrderedDict[str, CacheEntry] = OrderedDict()
self._lock = asyncio.Lock()
def _compute_key(self, messages: list[dict], model: str, **key_fields: Any) -> str:
def _compute_key(
self, messages: list[dict], model: str, *, partition: str, **key_fields: Any
) -> str:
"""Compute cache key from messages, model, and response-shaping fields.
``key_fields`` carries every request field that changes generation,
@@ -75,12 +77,19 @@ class SemanticCache:
on ``system``/``tools`` does not fragment the key (scalars pass through
untouched). Absent fields don't contribute, so truly-identical requests
still hit.
"""
return compute_semantic_cache_key(messages, model, **key_fields)
async def get(self, messages: list[dict], model: str, **key_fields: Any) -> CacheEntry | None:
``partition`` is the caller's cache partition
(:func:`headroom.proxy.semantic_cache_key_policy.compute_request_cache_partition`)
and is required: on a shared proxy, a response is only ever replayed to a
caller presenting the same provider credentials and principal.
"""
return compute_semantic_cache_key(messages, model, partition=partition, **key_fields)
async def get(
self, messages: list[dict], model: str, *, partition: str, **key_fields: Any
) -> CacheEntry | None:
"""Get cached response if exists and not expired."""
key = self._compute_key(messages, model, **key_fields)
key = self._compute_key(messages, model, partition=partition, **key_fields)
async with self._lock:
entry = self._cache.get(key)
@@ -105,10 +114,12 @@ class SemanticCache:
response_body: bytes,
response_headers: dict[str, str],
tokens_saved: int = 0,
*,
partition: str,
**key_fields: Any,
):
"""Cache a response."""
key = self._compute_key(messages, model, **key_fields)
key = self._compute_key(messages, model, partition=partition, **key_fields)
if not _is_cacheable_reply(response_body):
return
+103
View File
@@ -3,9 +3,104 @@
from __future__ import annotations
import hashlib
import hmac
import json
import logging
import re
import secrets
from collections.abc import Mapping
from typing import Any
from headroom.proxy.internal_header_policy import INTERNAL_HEADER_PREFIX, is_credential_header
logger = logging.getLogger(__name__)
# Per-process key for the cache partition HMAC. The semantic response cache is
# an in-memory, per-process structure, so the partition only has to be stable
# for this process's lifetime; a fresh random key means a partition id can
# never be computed (or brute-forced from a guessed credential) outside it.
_PARTITION_KEY = secrets.token_bytes(32)
# A header carries caller credentials or selects the billed account when the
# shared credential-header rule matches it (``authorization``,
# ``proxy-authorization``, ``cookie``, ``x-api-key``, ...) or its name ends in
# one of these tokens (``x-goog-api-key``, ``chatgpt-account-id``,
# ``openai-organization``, ``openai-project``, ...). Every such header the
# handlers forward upstream must be in the partition. Matching by shape rather
# than by a fixed provider list keeps new providers partitioned by default.
_CREDENTIAL_HEADER_RE = re.compile(
r"(^|[-_])(api[-_]?key|key|token|secret|account|account[-_]id|organization|project)$"
)
# Per-request nonces that match the credential shape but identify nothing: a
# fresh value per call would make every request a unique partition.
_NON_CREDENTIAL_HEADERS = frozenset({"idempotency-key", "x-idempotency-key"})
ANONYMOUS_PARTITION = "anon"
def _is_credential_header(name: str) -> bool:
# ``x-headroom-*`` headers (including the proxy's own token, which the
# security gate has already removed) never reach the upstream, so they
# never identify the upstream account.
lowered = name.lower()
if lowered.startswith(INTERNAL_HEADER_PREFIX) or lowered in _NON_CREDENTIAL_HEADERS:
return False
return is_credential_header(lowered) or bool(_CREDENTIAL_HEADER_RE.search(lowered))
def compute_cache_partition(
headers: Mapping[str, str] | Any,
*,
principal: str | None = None,
) -> str:
"""Return the response-cache partition for one caller.
Two requests may share a cached response only when they present the same
provider credentials / account selectors and the same authenticated
principal. The partition is an HMAC under a per-process random key, never a
plain hash, so it cannot be used to confirm a guessed credential.
Requests with no caller credential and no principal (the proxy supplies the
operator's key) share the ``anon`` partition: they are billed to, and
answered by, the same account.
"""
material: list[tuple[str, str]] = []
items: list[tuple[Any, Any]] = list(headers.items()) if hasattr(headers, "items") else []
for name, value in items:
if value and _is_credential_header(str(name)):
material.append(("h:" + str(name).lower(), str(value).strip()))
if principal:
material.append(("principal", principal))
if not material:
return ANONYMOUS_PARTITION
material.sort()
digest = hmac.new(
_PARTITION_KEY, json.dumps(material, separators=(",", ":")).encode(), hashlib.sha256
).hexdigest()
return "p_" + digest[:32]
def compute_request_cache_partition(request: Any) -> str | None:
"""Partition for a live proxy request: credentials + authenticated principal.
Returns ``None`` when an installed identity resolver raises or returns no
principal. The caller must then bypass the response cache (no lookup, no
store): falling back to the credential-only partition would let tenants
sharing one operator key read each other's cached responses.
"""
from headroom.proxy.identity import resolve_authenticated_principal
try:
principal = resolve_authenticated_principal(request)
except Exception:
logger.warning(
"Identity resolver could not establish a principal; "
"bypassing the response cache for this request",
exc_info=True,
)
return None
headers = getattr(request, "headers", None) or {}
return compute_cache_partition(headers, principal=principal)
def strip_cache_control(obj: Any) -> Any:
"""Recursively drop prompt-cache annotations, preserving schema properties."""
@@ -33,6 +128,8 @@ def _strip_cache_control(obj: Any, *, preserve_key: bool) -> Any:
def compute_semantic_cache_key(
messages: list[dict],
model: str,
*,
partition: str,
**key_fields: Any,
) -> str:
"""Compute a stable cache key from request content and shaping fields.
@@ -44,9 +141,15 @@ def compute_semantic_cache_key(
Anthropic path, the most common place a client (e.g. Claude Code) moves a
breakpoint between turns, so leaving them un-stripped defeated the strip for
the field that matters most.
``partition`` (see :func:`compute_cache_partition`) is mandatory: a shared
proxy must never answer one caller from another caller's cached response.
"""
if not partition:
raise ValueError("semantic cache key requires a caller partition")
normalized = json.dumps(
{
"partition": partition,
"model": model,
"messages": strip_cache_control(messages),
**{k: strip_cache_control(v) for k, v in key_fields.items()},
@@ -27,6 +27,7 @@ httpx = pytest.importorskip("httpx")
from fastapi.testclient import TestClient # noqa: E402
from headroom.proxy.models import CacheEntry # noqa: E402
from headroom.proxy.semantic_cache_key_policy import compute_cache_partition # noqa: E402
from headroom.proxy.server import ProxyConfig, create_app # noqa: E402
RETRIEVE_TOOL = {
@@ -314,6 +315,8 @@ def test_cache_hit_never_replays_a_foreign_content_type() -> None:
key = proxy.cache._compute_key(
body["messages"],
body["model"],
# Same partition the handler derives for this caller (01-F15).
partition=compute_cache_partition(_headers()),
upstream_base_url=None,
system=None,
tools=None,
@@ -0,0 +1,390 @@
"""01-F15: the response cache must never answer one caller from another's entry.
On a shared proxy (several principals behind one Headroom), two callers sending
the identical request under different provider credentials used to share one
semantic-cache entry: caller B was served the completion generated for caller A,
under A's key. The partition is now a keyed HMAC of the caller's credential
headers (and authenticated principal, when an identity resolver is installed),
threaded through every cache lookup and store.
"""
from __future__ import annotations
import hashlib
from types import SimpleNamespace
import httpx
import pytest
pytest.importorskip("fastapi")
from fastapi.testclient import TestClient
from headroom.proxy import identity
from headroom.proxy.semantic_cache import SemanticCache
from headroom.proxy.semantic_cache_key_policy import (
ANONYMOUS_PARTITION,
compute_cache_partition,
compute_request_cache_partition,
)
from headroom.proxy.server import ProxyConfig, create_app
# ---------------------------------------------------------------------------
# Partition function
# ---------------------------------------------------------------------------
def test_partition_differs_per_credential_and_is_stable_per_credential() -> None:
a = compute_cache_partition({"x-api-key": "sk-ant-A"})
b = compute_cache_partition({"x-api-key": "sk-ant-B"})
assert a != b
assert a == compute_cache_partition({"X-Api-Key": "sk-ant-A"})
@pytest.mark.parametrize(
"header",
[
"authorization",
"proxy-authorization",
"cookie",
"x-api-key",
"api-key",
"x-goog-api-key",
"chatgpt-account-id",
"openai-organization",
"openai-project",
],
)
def test_every_credential_or_account_header_partitions(header: str) -> None:
assert compute_cache_partition({header: "one"}) != compute_cache_partition({header: "two"})
def test_non_credential_headers_do_not_fragment_the_cache() -> None:
base = compute_cache_partition({"x-api-key": "k"})
noisy = {
"x-api-key": "k",
"idempotency-key": "per-request-nonce",
"x-headroom-proxy-token": "proxy-secret",
"user-agent": "sdk/1.0",
"anthropic-version": "2023-06-01",
"content-type": "application/json",
}
assert compute_cache_partition(noisy) == base
def test_no_credential_and_no_principal_is_the_shared_operator_partition() -> None:
assert compute_cache_partition({"user-agent": "x"}) == ANONYMOUS_PARTITION
def test_partition_is_keyed_not_a_plain_hash_of_the_credential() -> None:
part = compute_cache_partition({"x-api-key": "sk-ant-A"})
for plain in (
hashlib.sha256(b"sk-ant-A").hexdigest()[:32],
hashlib.sha256(b"x-api-key:sk-ant-A").hexdigest()[:32],
):
assert plain not in part
def test_authenticated_principal_partitions_callers_sharing_a_credential() -> None:
same_key = {"authorization": "Bearer operator-key"}
assert compute_cache_partition(same_key, principal="alice") != compute_cache_partition(
same_key, principal="bob"
)
def test_semantic_cache_refuses_calls_without_a_partition() -> None:
cache = SemanticCache()
with pytest.raises(TypeError):
cache._compute_key([{"role": "user", "content": "x"}], "m") # type: ignore[call-arg]
@pytest.mark.asyncio
async def test_semantic_cache_isolates_partitions() -> None:
cache = SemanticCache()
msgs = [{"role": "user", "content": "Say hi."}]
await cache.set(msgs, "m", b"A-body", {}, partition="p_a")
assert await cache.get(msgs, "m", partition="p_b") is None
hit = await cache.get(msgs, "m", partition="p_a")
assert hit is not None and hit.response_body == b"A-body"
# ---------------------------------------------------------------------------
# End to end through the real handlers
# ---------------------------------------------------------------------------
def _client(*, cache_enabled: bool = True) -> TestClient:
return TestClient(
create_app(
ProxyConfig(
optimize=False,
cache_enabled=cache_enabled,
rate_limit_enabled=False,
cost_tracking_enabled=False,
log_requests=False,
ccr_inject_tool=False,
ccr_handle_responses=False,
ccr_context_tracking=False,
image_optimize=False,
)
)
)
def _anthropic_reply(text: str) -> httpx.Response:
return httpx.Response(
200,
json={
"id": "msg_1",
"type": "message",
"role": "assistant",
"content": [{"type": "text", "text": text}],
"usage": {
"input_tokens": 10,
"output_tokens": 3,
"cache_read_input_tokens": 0,
"cache_creation_input_tokens": 0,
},
},
)
def _openai_reply(text: str) -> httpx.Response:
return httpx.Response(
200,
json={
"id": "chatcmpl_1",
"object": "chat.completion",
"choices": [
{
"index": 0,
"message": {"role": "assistant", "content": text},
"finish_reason": "stop",
}
],
"usage": {"prompt_tokens": 10, "completion_tokens": 3, "total_tokens": 13},
},
)
ANTHROPIC_BODY = {
"model": "claude-haiku-4-5",
"max_tokens": 64,
"messages": [{"role": "user", "content": "What is my account's secret plan?"}],
"stream": False,
}
OPENAI_BODY = {
"model": "gpt-4o-mini",
"max_tokens": 64,
"messages": [{"role": "user", "content": "What is my account's secret plan?"}],
"stream": False,
}
def _drive(path: str, body: dict, reply, header: str, fmt: str): # noqa: ANN001
"""Send A, B (different credential), A again; return bodies and upstream owners."""
seen: list[str] = []
out: list[str] = []
with _client() as client:
proxy = client.app.state.proxy
async def _fake_retry(method, url, headers, body, stream=False, **kwargs): # noqa: ANN001
lowered = {k.lower(): v for k, v in headers.items()}
owner = "A" if "key-A" in lowered.get(header, "") else "B"
seen.append(owner)
return reply(f"private answer for {owner}")
proxy._retry_request = _fake_retry
extra = {"anthropic-version": "2023-06-01"} if path == "/v1/messages" else {}
for who in ("A", "B", "A"):
r = client.post(path, headers={header: fmt.format(who), **extra}, json=body)
assert r.status_code == 200, r.text
out.append(r.text)
return out, seen
@pytest.mark.parametrize(
"path,body,reply,header,fmt",
[
("/v1/messages", ANTHROPIC_BODY, _anthropic_reply, "x-api-key", "key-{}"),
("/v1/chat/completions", OPENAI_BODY, _openai_reply, "authorization", "Bearer key-{}"),
# Cookie and Proxy-Authorization are forwarded upstream, so a
# cookie-authenticated gateway must not let two sessions share a reply.
("/v1/messages", ANTHROPIC_BODY, _anthropic_reply, "cookie", "session=key-{}"),
("/v1/chat/completions", OPENAI_BODY, _openai_reply, "cookie", "session=key-{}"),
(
"/v1/messages",
ANTHROPIC_BODY,
_anthropic_reply,
"proxy-authorization",
"Basic key-{}",
),
(
"/v1/chat/completions",
OPENAI_BODY,
_openai_reply,
"proxy-authorization",
"Basic key-{}",
),
],
ids=[
"anthropic",
"openai",
"anthropic-cookie",
"openai-cookie",
"anthropic-proxy-authorization",
"openai-proxy-authorization",
],
)
def test_second_caller_is_never_served_first_callers_cached_response(
path, body, reply, header, fmt
) -> None:
(a, b, a_again), upstream = _drive(path, body, reply, header, fmt)
# B must reach the upstream under its own key and never see A's answer.
assert "private answer for A" not in b
assert "private answer for B" in b
# A's repeat is still a cache hit: single-credential hit rate is unchanged.
assert "private answer for A" in a_again
assert upstream == ["A", "B"]
def test_installed_identity_resolver_partitions_callers_on_one_operator_key() -> None:
identity.set_identity_resolver(
lambda request, *, default: request.headers.get("x-test-user", "")
)
try:
with _client() as client:
proxy = client.app.state.proxy
calls: list[str] = []
async def _fake_retry(method, url, headers, body, stream=False, **kwargs): # noqa: ANN001
calls.append("x")
return _anthropic_reply(f"answer {len(calls)}")
proxy._retry_request = _fake_retry
common = {"x-api-key": "shared-operator-key", "anthropic-version": "2023-06-01"}
r1 = client.post(
"/v1/messages", headers={**common, "x-test-user": "alice"}, json=ANTHROPIC_BODY
)
r2 = client.post(
"/v1/messages", headers={**common, "x-test-user": "bob"}, json=ANTHROPIC_BODY
)
assert "answer 1" in r1.text
assert "answer 2" in r2.text
assert len(calls) == 2
finally:
identity.set_identity_resolver(None)
def _failing_resolver(request, *, default): # noqa: ANN001
raise RuntimeError("identity backend unavailable")
def _empty_resolver(request, *, default): # noqa: ANN001
# Installed, but cannot establish a principal for this caller.
return ""
UNRESOLVED_RESOLVERS = pytest.mark.parametrize(
"resolver", [_failing_resolver, _empty_resolver], ids=["raises", "empty"]
)
@UNRESOLVED_RESOLVERS
def test_failing_identity_resolver_yields_no_partition(resolver) -> None: # noqa: ANN001
identity.set_identity_resolver(resolver)
try:
request = SimpleNamespace(headers={"x-api-key": "shared-operator-key"})
assert compute_request_cache_partition(request) is None
finally:
identity.set_identity_resolver(None)
def test_no_installed_resolver_keeps_the_credential_partition() -> None:
request = SimpleNamespace(headers={"x-api-key": "shared-operator-key"})
assert compute_request_cache_partition(request) == compute_cache_partition(request.headers)
@pytest.mark.parametrize(
"path,body,reply,headers",
[
(
"/v1/messages",
ANTHROPIC_BODY,
_anthropic_reply,
{"x-api-key": "shared-operator-key", "anthropic-version": "2023-06-01"},
),
(
"/v1/chat/completions",
OPENAI_BODY,
_openai_reply,
{"authorization": "Bearer shared-operator-key"},
),
],
ids=["anthropic", "openai"],
)
@UNRESOLVED_RESOLVERS
def test_failing_identity_resolver_bypasses_the_response_cache(
resolver, # noqa: ANN001
path, # noqa: ANN001
body, # noqa: ANN001
reply, # noqa: ANN001
headers, # noqa: ANN001
) -> None:
# A resolver that raises or returns no principal must not collapse tenants
# on one operator key into the credential-only partition: the cache is
# neither read nor written.
identity.set_identity_resolver(resolver)
try:
with _client() as client:
proxy = client.app.state.proxy
calls: list[str] = []
async def _fake_retry(method, url, headers, body, stream=False, **kwargs): # noqa: ANN001
calls.append("x")
return reply(f"answer {len(calls)}")
proxy._retry_request = _fake_retry
texts = [
client.post(path, headers={**headers, "x-test-user": who}, json=body).text
for who in ("alice", "bob", "alice")
]
assert "answer 1" in texts[0]
assert "answer 1" not in texts[1] and "answer 2" in texts[1]
assert "answer 3" in texts[2]
assert len(calls) == 3
finally:
identity.set_identity_resolver(None)
@pytest.mark.parametrize(
"path,body,reply,headers",
[
("/v1/messages", ANTHROPIC_BODY, _anthropic_reply, {"x-api-key": "k"}),
("/v1/chat/completions", OPENAI_BODY, _openai_reply, {"authorization": "Bearer k"}),
],
ids=["anthropic", "openai"],
)
def test_cache_disabled_requests_skip_identity_resolution(path, body, reply, headers) -> None:
# No response cache means no partition is needed, so the (possibly
# remote) identity resolver must not be consulted for it.
resolved: list[str] = []
def _counting_resolver(request, *, default): # noqa: ANN001
resolved.append("x")
return "alice"
identity.set_identity_resolver(_counting_resolver)
try:
with _client(cache_enabled=False) as client:
proxy = client.app.state.proxy
async def _fake_retry(method, url, headers, body, stream=False, **kwargs): # noqa: ANN001
return reply("ok")
proxy._retry_request = _fake_retry
extra = {"anthropic-version": "2023-06-01"} if path == "/v1/messages" else {}
assert client.post(path, headers={**headers, **extra}, json=body).status_code == 200
assert resolved == []
finally:
identity.set_identity_resolver(None)
+11 -7
View File
@@ -22,7 +22,7 @@ MODEL = "claude-haiku-4-5"
def _key(cache: SemanticCache, **kw) -> str:
return cache._compute_key(MESSAGES, MODEL, **kw)
return cache._compute_key(MESSAGES, MODEL, partition="p_test", **kw)
# --- too loose: different inputs must NOT collide (the bug) --------------------
@@ -145,8 +145,8 @@ def test_message_cache_control_breakpoint_move_same_key():
}
]
messages_without_cc = [{"role": "user", "content": [{"type": "text", "text": "hello"}]}]
assert cache._compute_key(messages_with_cc, MODEL) == cache._compute_key(
messages_without_cc, MODEL
assert cache._compute_key(messages_with_cc, MODEL, partition="p_test") == cache._compute_key(
messages_without_cc, MODEL, partition="p_test"
)
@@ -155,7 +155,9 @@ def test_message_content_change_still_distinct_key():
cache = SemanticCache()
a = [{"role": "user", "content": [{"type": "text", "text": "hello"}]}]
b = [{"role": "user", "content": [{"type": "text", "text": "goodbye"}]}]
assert cache._compute_key(a, MODEL) != cache._compute_key(b, MODEL)
assert cache._compute_key(a, MODEL, partition="p_test") != cache._compute_key(
b, MODEL, partition="p_test"
)
# --- behavioral get/set: collision prevented end to end -----------------------
@@ -165,11 +167,13 @@ async def test_get_set_collision_prevented():
"""Store under system A; fetching with system B is a MISS (no contamination),
fetching with system A is a HIT."""
cache = SemanticCache()
await cache.set(MESSAGES, MODEL, b"french-body", {}, system="Answer only in French.")
await cache.set(
MESSAGES, MODEL, b"french-body", {}, partition="p_test", system="Answer only in French."
)
miss = await cache.get(MESSAGES, MODEL, system="Answer only in English.")
miss = await cache.get(MESSAGES, MODEL, partition="p_test", system="Answer only in English.")
assert miss is None
hit = await cache.get(MESSAGES, MODEL, system="Answer only in French.")
hit = await cache.get(MESSAGES, MODEL, partition="p_test", system="Answer only in French.")
assert hit is not None
assert hit.response_body == b"french-body"
+12 -8
View File
@@ -12,11 +12,11 @@ MODEL = "claude-haiku-4-5"
def test_semantic_cache_key_distinguishes_response_shaping_fields() -> None:
assert compute_semantic_cache_key(MESSAGES, MODEL, system="French") != (
compute_semantic_cache_key(MESSAGES, MODEL, system="English")
assert compute_semantic_cache_key(MESSAGES, MODEL, partition="p_test", system="French") != (
compute_semantic_cache_key(MESSAGES, MODEL, partition="p_test", system="English")
)
assert compute_semantic_cache_key(MESSAGES, MODEL, temperature=0.0) != (
compute_semantic_cache_key(MESSAGES, MODEL, temperature=1.0)
assert compute_semantic_cache_key(MESSAGES, MODEL, partition="p_test", temperature=0.0) != (
compute_semantic_cache_key(MESSAGES, MODEL, partition="p_test", temperature=1.0)
)
@@ -24,8 +24,12 @@ def test_semantic_cache_key_ignores_moved_cache_control() -> None:
with_cache_control = [{"type": "text", "text": "sys", "cache_control": {"type": "ephemeral"}}]
without_cache_control = [{"type": "text", "text": "sys"}]
assert compute_semantic_cache_key(MESSAGES, MODEL, system=with_cache_control) == (
compute_semantic_cache_key(MESSAGES, MODEL, system=without_cache_control)
assert compute_semantic_cache_key(
MESSAGES, MODEL, partition="p_test", system=with_cache_control
) == (
compute_semantic_cache_key(
MESSAGES, MODEL, partition="p_test", system=without_cache_control
)
)
@@ -38,8 +42,8 @@ def test_semantic_cache_key_ignores_moved_message_cache_control() -> None:
}
]
messages_without_cc = [{"role": "user", "content": [{"type": "text", "text": "hello"}]}]
assert compute_semantic_cache_key(messages_with_cc, MODEL) == (
compute_semantic_cache_key(messages_without_cc, MODEL)
assert compute_semantic_cache_key(messages_with_cc, MODEL, partition="p_test") == (
compute_semantic_cache_key(messages_without_cc, MODEL, partition="p_test")
)
+12 -3
View File
@@ -30,19 +30,23 @@ def test_strip_cache_control_recurses_through_dicts_and_lists() -> None:
def test_compute_semantic_cache_key_is_stable_for_identical_inputs() -> None:
kwargs = {"system": "sys", "tools": [{"name": "read"}], "temperature": 0.2}
assert compute_semantic_cache_key(MESSAGES, MODEL, **kwargs) == compute_semantic_cache_key(
assert compute_semantic_cache_key(
MESSAGES, MODEL, partition="p_test", **kwargs
) == compute_semantic_cache_key(
MESSAGES,
MODEL,
partition="p_test",
**kwargs,
)
def test_compute_semantic_cache_key_distinguishes_response_shaping_fields() -> None:
assert compute_semantic_cache_key(
MESSAGES, MODEL, temperature=0.0
MESSAGES, MODEL, partition="p_test", temperature=0.0
) != compute_semantic_cache_key(
MESSAGES,
MODEL,
partition="p_test",
temperature=1.0,
)
@@ -53,10 +57,12 @@ def test_compute_semantic_cache_key_ignores_moved_cache_control_breakpoints() ->
assert compute_semantic_cache_key(
MESSAGES,
MODEL,
partition="p_test",
system=with_breakpoint,
) == compute_semantic_cache_key(
MESSAGES,
MODEL,
partition="p_test",
system=without_breakpoint,
)
@@ -64,8 +70,11 @@ def test_compute_semantic_cache_key_ignores_moved_cache_control_breakpoints() ->
def test_semantic_cache_private_key_wrapper_delegates_to_policy() -> None:
cache = SemanticCache()
kwargs = {"system": "sys", "tools": [{"name": "read"}], "temperature": 0.2}
assert cache._compute_key(MESSAGES, MODEL, **kwargs) == compute_semantic_cache_key(
assert cache._compute_key(
MESSAGES, MODEL, partition="p_test", **kwargs
) == compute_semantic_cache_key(
MESSAGES,
MODEL,
partition="p_test",
**kwargs,
)