procurement-ingest/lambdas/wo/shoc_emitter/delivery.py
Adam Moussa c040050373
feat(webhook): SHOC WO webhook emitter - dark-ship streams + HMAC secret/rotation (PR-2) (#137)
* docs(webhook): revise SHOC webhook contract and plan for post-migration reality

Branch re-cut on main 2026-07-23 (old base carried stale PR #99 commits).

Contract Rev 2026-07-23:
- Producer account corrected: seahaven-prod (011934824531); mgmt frozen
- Reconciliation backstop is the new procurement read API, not SyncController
- wo_status "unknown" is real; SHOC must map it (checklist item added)
- write_origin forward-compat note for phase-2 write-back echo suppression
- SyncVendorReplies retirement flagged (dead table, no vendor_reply event)

Plan updates:
- Account gate: seahaven-prod only; never enable streams on mgmt tables
- Emitter ships DARK (ESMs enabled=False); activation is a deliberate flip
  after the SHOC receiver passes shared HMAC vectors
- Post-refactor conventions: common.py helpers, bundle-consistency AST pins,
  pytest.ini --cov additions, consolidated test roots
- Dedicated-CMK rationale, secret-ARN handooff step, consumer audit refreshed
  (slack-bot decommissioned), enum golden test, write_origin skip-branch test

* feat(webhook): SHOC WO webhook emitter — dark-ship streams, HMAC secret + rotation

Implements docs/shoc-webhook-plan.md Phases 1-5 (PR-2 of the SHOC
call-and-be-called effort). Everything ships DARK: both DynamoDB event
source mappings deploy enabled=False; activation is a deliberate
one-line follow-up PR gated on the SHOC receiver passing the shared
HMAC test vectors.

- Streams: NEW_AND_OLD_IMAGES on WorkOrders + WorkOrderComments
  (in-place update, RETAIN + logical IDs untouched; no existing
  consumers — verified live, neither table had a stream).
- workorder-shoc-emitter (Py3.12/ARM64): stream -> envelope ->
  HMAC-signed POST per docs/shoc-webhook-contract.md; strict per-shard
  ordering (parallelization 1, bisect off, retry until 24h age,
  ReportBatchItemFailures); 429/5xx/timeout block the shard in order,
  other 4xx park to workorder-shoc-emitter-rejected; ESM failures ->
  workorder-shoc-emitter-failures (metadata; replay rebuilds from
  DynamoDB). Echo guard skips write_origin=shoc-write-api.
- Secret workorder-ingest/shoc-webhook-hmac on a dedicated CMK
  (alias workorder-ingest-shoc-webhook-kms); cross-account
  GetSecretValue/DescribeSecret + kms:Decrypt granted to exactly
  arn:aws:iam::396287094661:role/shoc-backend-dev. RemovalPolicy
  DESTROY deliberately (machine-generated material; avoids the
  fixed-name RETAIN-orphan deadlock).
- workorder-shoc-hmac-rotator: 30-day rotation, dual-key overlap,
  64-hex keys, kid = UTC %Y-%m-%dT%H.
- Alarms (ALARM-only -> site-alerts): emitter errors/throttles/
  duration + iterator-age (>=10 min) + failures/rejected queue
  depth; rotator standard trio.
- scripts/replay_shoc_webhooks.py: dry-run-default operator replay
  (rebuilds from tables, replay:true envelopes).
- Tests: 742 passing, 85.56% aggregate; golden HMAC vectors shared
  with SHOC in docs/shoc-webhook-test-vectors.json (emitter + replay
  signing pinned to identical vectors); bundle-consistency AST pins
  for both new bundles.
- README: WO stack + webhook feed section, alarm table, runbooks;
  removed stale seahaven-slack-bot consumer references.

* fix(webhook): kms:ViaService pins, https-only delivery, cross-account principal CI pin

GPT-4.1 cross-family review of the policy surface (no BLOCK): FIX applied
to the cross-account shoc-backend-dev Decrypt statement and both Lambda
role KMS grants (the key is only ever used via Secrets Manager); its
invariant-enforcement QUESTION answered durably with
tests/test_cross_account_principal_pin.py (any new foreign IAM principal
in cdk/ fails CI). Scanner mediums fixed: delivery.py and the replay
script now refuse non-https URLs (urllib follows file:// and http://).
SQS metadata-action and dynamodb:ListStreams NITs skipped: standard CDK
grant shapes; ListStreams has no resource-level scoping. The 4 gitleaks
HIGHs on docs/shoc-webhook-test-vectors.json are deliberate non-secrets
(shared receiver-verification vectors) suppressed machine-level with
justification.

* harden(webhook): resolve /sh-security-review findings (1 confirmed medium + cheap fixes)

High-recall detector fan-out (injection/authz/secrets-crypto/iac-iam/logic)
+ proof-or-kill verifier. Gate PASSES: 1 confirmed medium, 0 confirmed
critical/high. Confirmed finding fixed; several unverified-but-cheap
hardenings applied since the emitter ships dark and activation is weeks out.

- CONFIRMED medium (confused deputy): the rotation Lambda's generated
  invoke permission for secretsmanager.amazonaws.com carried no
  SourceAccount/SourceArn, so any account's Secrets Manager could invoke
  the rotator. Patched the generated CfnPermission in place (a second
  permission would be additive, not restrictive) to pin account + this
  secret ARN.
- delivery + replay: refuse to follow receiver 3xx redirects (no-redirect
  opener) so live X-SH-* auth headers can't be forwarded to a
  receiver-chosen Location and an http:// Location can't slip past the
  https guard. Fixed the "unfollowed 3xx" comment that was factually wrong.
- delivery: classify 401/403 as retryable (invalidate key cache + retry in
  order) instead of parking -- transient auth failures (rotation outran the
  TTL cache, clock skew) are availability events, not contract bugs.
- envelope: build_event now genuinely total (guarded eventID /
  ApproximateCreationDateTime subscripts) per its own never-raise contract.
- handler: catch-all so an unexpected per-record error (e.g. SQS park
  failure) reports only that record instead of failing the whole batch
  (which would re-deliver every earlier success for 24h); per-invocation
  emit/skip batch summary so a systemic silent drop is queryable/alarmable.
- rotator: narrow the AWSCURRENT-read except to ResourceNotFound/JSONDecode
  (transient SM/KMS errors re-raise so the overlap key isn't silently
  dropped); kid uniqueness checked against ALL retained kids with a random
  suffix on collision (never reissue a kid for a different secret).
- contract: skeleton-upsert required on ANY unknown work_order_id (not just
  comment-before-create) + monotonicity guard (ignore older updated_at), so
  a parked created or an out-of-order replay can't corrupt receiver state.

Unverified/refuted findings left as-is with rationale: the two "high" logic
claims (whole-batch crash triggers, ordering violation) were refuted on
reachability (real stream records carry required fields; persistence writes
strings only; full-state idempotent upsert absorbs the ordering gap). Signed
kid/version binding (AUTHZ-002) declined: coordinated contract change, not
cheap, no exploit with one algorithm/key.

* fix(webhook): drop kid from rotator test_ok log (CodeQL clear-text-logging FP)

GHAS CodeQL flagged py/clear-text-logging-sensitive-data (high) at
_test_secret's success log because head["kid"] is subscripted from the
same parsed-secret dict that holds head["secret"] — the taint tracker
can't tell the non-secret key id from the secret. The secret value is
never logged. Rather than dismiss the alert (fragile; re-alerts on line
moves), remove the flow: kid is already logged at stage time in
_create_secret and version_id correlates the steps, so the test_ok log
keeps only event + version_id. Also hardens against a future edit that
swaps the logged field.
2026-07-24 22:12:20 +00:00

218 lines
9.1 KiB
Python

"""HMAC-signed webhook delivery to the SHOC receiver (contract sections 6-7).
Owns the signed POST and its response classification. The signing key comes
from Secrets Manager (``workorder-ingest/shoc-webhook-hmac``, dual-key shape
``{"keys": [{"kid", "secret"}, ...]}``), cached in the warm container for a
short TTL so 30-day rotation propagates without waiting for the execution
environment to recycle -- the producer always signs with ``keys[0]``.
Response contract (section 7): 2xx -> delivered; 429/5xx/timeout/connection
error -> RetryableDeliveryError (the handler surfaces it as a batch item
failure so the ESM blocks the shard and retries in order); any other status ->
("rejected", code) for the handler to park (a contract bug must not block the
shard for 24 hours). Secret material and signatures are never logged.
``sign_body`` is a module-level pure function on purpose: the golden-vector
test suite and the replay script both pin against it, and it is the shared
definition Luby's receiver verifies with.
"""
import hashlib
import hmac
import json
import logging
import os
import time
import urllib.error
import urllib.request
from http import HTTPStatus
import boto3
from botocore.config import Config
logger = logging.getLogger()
logger.setLevel(logging.INFO)
# Required, NO default (Open SWE #0): a hardcoded fallback URL would silently
# ship production work-order state to that endpoint if the CDK-set env var were
# ever dropped. Unset -> deliver() fails closed (retries in order) rather than
# defaulting anywhere. The single source of the URL is the Lambda environment
# (CDK), confirmed by the activation PR.
SHOC_WEBHOOK_URL = os.environ.get("SHOC_WEBHOOK_URL")
HMAC_SECRET_ARN = os.environ.get("HMAC_SECRET_ARN")
POST_TIMEOUT_SECONDS = 10
USER_AGENT = "workorder-shoc-emitter/1"
# Bound the Secrets Manager client's timeouts (Open SWE #16): default botocore
# timeouts can run to ~60s, which combined with the 10s POST could exhaust the
# Lambda budget and make a slow AWS call re-deliver the whole batch. Fast, few
# retries -- the key is 300s-cached so this fetch is rare.
_BOTO_CONFIG = Config(
connect_timeout=3, read_timeout=5, retries={"max_attempts": 2, "mode": "standard"}
)
# 2xx = delivered; 429 and 5xx retry; everything else parks (contract sec. 7).
_HTTP_SUCCESS_RANGE = range(HTTPStatus.OK, HTTPStatus.MULTIPLE_CHOICES)
_HTTP_SERVER_ERROR_MIN = HTTPStatus.INTERNAL_SERVER_ERROR
# Transient auth failures: a stale cached key mid-rotation, a receiver
# secret-fetch blip, or clock skew past the +/-300s window. These are
# availability events, not payload contract bugs, so they retry (in order)
# after the key cache is dropped -- never park, which would strand the
# delivery out of the ordered feed until a manual replay.
_HTTP_AUTH_FAILURES = (HTTPStatus.UNAUTHORIZED, HTTPStatus.FORBIDDEN)
class _NoRedirectHandler(urllib.request.HTTPRedirectHandler):
"""Refuse to follow receiver redirects.
The default opener follows 3xx transparently, which would (a) forward the
live X-SH-* auth headers to a receiver-chosen Location and (b) let an
http:// Location slip past the https-only guard on the configured URL. We
only ever POST to the one configured endpoint; a redirect is a receiver
misconfiguration, so surface the 3xx as an HTTPError and let it park.
"""
def redirect_request(self, *args, **kwargs):
return None
_opener = urllib.request.build_opener(_NoRedirectHandler)
# Refresh the cached key material this often so a rotated secret propagates
# without waiting for the execution environment to recycle (contract requires
# receiver-side TTL <= 300s; the producer matches it).
_HMAC_KEYS_CACHE_TTL_SECONDS = 300
_hmac_keys_cache = None
_hmac_keys_cached_at = 0.0
class RetryableDeliveryError(Exception):
"""Delivery failed in a way the ESM should retry in order (shard-blocking).
``status_code`` carries the HTTP status when one exists (429/5xx); it is
None for timeouts, connection errors, and secret-fetch failures.
"""
def __init__(self, message: str, status_code: int | None = None):
super().__init__(message)
self.status_code = status_code
def _get_hmac_keys() -> list[dict]:
"""Fetch the signing-key list from Secrets Manager, TTL-cached.
An empty/missing ``keys`` list is the bootstrap state before the first
rotation has run -- retryable, not a crash: the shard blocks until the
rotator populates the secret. Fetch failures (throttle, transient IAM/KMS
denial) are likewise retryable so a Secrets Manager blip blocks in order
instead of failing the whole batch. Key material is never logged.
"""
global _hmac_keys_cache, _hmac_keys_cached_at
now = time.monotonic()
if (
_hmac_keys_cache is not None
and now - _hmac_keys_cached_at < _HMAC_KEYS_CACHE_TTL_SECONDS
):
return _hmac_keys_cache
return _refresh_hmac_keys()
def _invalidate_hmac_keys() -> None:
"""Drop the cached key material so the next sign re-fetches (rotation)."""
global _hmac_keys_cache
_hmac_keys_cache = None
def _refresh_hmac_keys() -> list[dict]:
"""Force a fresh Secrets Manager fetch, bypassing the TTL cache."""
global _hmac_keys_cache, _hmac_keys_cached_at
now = time.monotonic()
secrets = boto3.client("secretsmanager", config=_BOTO_CONFIG)
try:
secret = secrets.get_secret_value(SecretId=HMAC_SECRET_ARN)
keys = json.loads(secret["SecretString"]).get("keys") or []
except Exception as exc:
raise RetryableDeliveryError(f"hmac secret fetch failed: {exc}") from exc
if not keys:
raise RetryableDeliveryError("hmac secret not yet rotated")
_hmac_keys_cache = keys
_hmac_keys_cached_at = now
return keys
def sign_body(secret_hex: str, timestamp: int, raw_body: bytes) -> str:
"""HMAC-SHA256 hex digest over ``f"{timestamp}.{raw_body}"`` (raw bytes).
The key is the UTF-8 bytes of the secret string exactly as stored in the
secret's ``secret`` field (the receiver reads the same JSON field -- no
hex-decoding on either side). Pure function; the golden-vector tests and
Luby's receiver both pin against this definition.
"""
string_to_sign = f"{timestamp}.".encode() + raw_body
return hmac.new(
secret_hex.encode("utf-8"), string_to_sign, hashlib.sha256
).hexdigest()
def deliver(envelope: dict) -> tuple[str, int]:
"""POST one envelope to SHOC. Returns ("delivered"|"rejected", status).
Raises RetryableDeliveryError for 429/5xx/timeout/connection failures so
the caller can block the shard (in-order retry, contract section 7).
"""
if not SHOC_WEBHOOK_URL or not SHOC_WEBHOOK_URL.startswith("https://"):
# Fail closed if the URL is unset (no hardcoded fallback -- Open SWE
# #0) or is any non-HTTPS scheme (file://, http://, ...): urllib would
# otherwise follow it, and the HMAC only protects an HTTPS transport.
# Retryable, so a config gap blocks the shard (visible via iterator-age)
# rather than shipping data somewhere unintended.
raise RetryableDeliveryError(
"SHOC_WEBHOOK_URL is unset or not an https:// URL; refusing to deliver"
)
raw_body = json.dumps(envelope).encode("utf-8")
timestamp = int(time.time())
signing_key = _get_hmac_keys()[0]
signature = sign_body(signing_key["secret"], timestamp, raw_body)
request = urllib.request.Request(
SHOC_WEBHOOK_URL,
data=raw_body,
headers={
"Content-Type": "application/json; charset=utf-8",
"User-Agent": USER_AGENT,
"X-SH-Timestamp": str(timestamp),
"X-SH-Key-Id": signing_key["kid"],
"X-SH-Signature": f"v1={signature}",
},
method="POST",
)
try:
# _opener refuses redirects, so a 3xx surfaces here as an HTTPError
# rather than silently re-issuing the request (with its auth headers)
# to a receiver-chosen Location.
with _opener.open(request, timeout=POST_TIMEOUT_SECONDS) as response:
status_code = response.status
except urllib.error.HTTPError as exc:
status_code = exc.code
if (
status_code == HTTPStatus.TOO_MANY_REQUESTS
or status_code >= _HTTP_SERVER_ERROR_MIN
):
raise RetryableDeliveryError(
f"receiver returned {status_code}", status_code=status_code
) from exc
if status_code in _HTTP_AUTH_FAILURES:
# Drop the cached key so an in-order retry re-signs with the
# current secret (handles a rotation that outran the TTL cache).
_invalidate_hmac_keys()
raise RetryableDeliveryError(
f"receiver auth failure {status_code}", status_code=status_code
) from exc
return ("rejected", status_code)
except (TimeoutError, urllib.error.URLError, OSError) as exc:
raise RetryableDeliveryError(f"connection error: {exc}") from exc
if status_code in _HTTP_SUCCESS_RANGE:
return ("delivered", status_code)
# A non-2xx that the opener did not raise for -- contract bug on
# someone's side. Park it, never block the shard.
return ("rejected", status_code)