Fix worker/gateway logic bugs and duplicate CORS headers
worker.py (both the ConfigMap copy and conf/worker.py): - Set an explicit ack_wait=60s on the JetStream pull consumer. It was previously left at the implicit default (~30s), the same ballpark as the outbound HTTP timeout - a slow-but-legitimate external call could cause JetStream to redeliver the message to another worker while the first was still mid-request, double-processing a non-idempotent call (e.g. duplicate order creation). - Track in-flight tasks and drain them (bounded wait) before closing the NATS/HTTP connections on shutdown, instead of cutting them off immediately - avoids dropped/duplicated messages on pod restarts. - Generic exception handler now does nak(delay=5) instead of an undelayed nak(), avoiding a tight redelivery loop on a persistent bug. - Missing 'data' field in a message now explicitly drops with a log line instead of silently forwarding the entire internal envelope. - Removed the hardcoded NATS password fallback baked into the source (every deployment already supplies it via a Secret at runtime, so this was a redundant plaintext copy sitting in a ConfigMap). app.py (both the ConfigMap copy and conf/app.py): - Fixed "NATS by connected" typo -> "NATS not connected". - Same hardcoded-password-fallback removal as worker.py. CORS: - conf/nginx-jupiter.conf and the in-cluster jupiter-cors-proxy nginx config both add their own CORS headers without stripping any the upstream might set, unlike nginx-queue-proxy.conf which does this correctly. Added proxy_hide_header for the ACA-* headers in both - browsers reject a response with duplicate Access-Control-* values. docker-compose.yml: - Added the missing doormile-proxy service (doormile.com -> :8206 -> NodePort 30830). nginx-doormile.conf existed but had no service wiring it into Traefik, unlike every other app.
This commit is contained in:
@@ -60,7 +60,7 @@ js = None
|
||||
# Configuration
|
||||
NATS_URL = os.getenv("NATS_URL", "nats://nats-server:4222")
|
||||
NATS_USER = os.getenv("NATS_USER", "admin")
|
||||
NATS_PASSWORD = os.getenv("NATS_PASSWORD", "package@321#")
|
||||
NATS_PASSWORD = os.getenv("NATS_PASSWORD", "")
|
||||
|
||||
# Endpoint to NATS subject mapping
|
||||
ENDPOINT_ROUTES = {
|
||||
@@ -172,7 +172,7 @@ async def publish_request_to_nats(endpoint: str, data: Dict[str, Any], request_m
|
||||
subject = ENDPOINT_ROUTES.get(endpoint, "api.unknown")
|
||||
|
||||
if not js:
|
||||
raise HTTPException(status_code=503, detail="NATS by connected")
|
||||
raise HTTPException(status_code=503, detail="NATS not connected")
|
||||
|
||||
try:
|
||||
# Create a unique inbox for the reply
|
||||
|
||||
@@ -12,7 +12,15 @@ http {
|
||||
# Proxy to Kubernetes NodePort (30822)
|
||||
# NodePorts are bound to 0.0.0.0 and are more reliable to access from host.docker.internal
|
||||
proxy_pass http://jupiter_k8s;
|
||||
|
||||
|
||||
# Strip any CORS headers the backend may set itself, so we don't
|
||||
# end up sending duplicate Access-Control-* headers (browsers
|
||||
# reject a response that has more than one value for these).
|
||||
proxy_hide_header 'Access-Control-Allow-Origin';
|
||||
proxy_hide_header 'Access-Control-Allow-Methods';
|
||||
proxy_hide_header 'Access-Control-Allow-Headers';
|
||||
proxy_hide_header 'Access-Control-Max-Age';
|
||||
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Real-IP $remote_addr;
|
||||
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
|
||||
|
||||
@@ -12,6 +12,7 @@ import asyncio
|
||||
import aiohttp
|
||||
import nats
|
||||
from nats.errors import TimeoutError as NatsTimeoutError
|
||||
from nats.js.api import ConsumerConfig
|
||||
from prometheus_client import Counter, Histogram, Gauge, start_http_server
|
||||
import signal
|
||||
import sys
|
||||
@@ -25,7 +26,7 @@ messages_in_flight = Gauge('worker_messages_in_flight', 'Messages currently bein
|
||||
# Configuration
|
||||
NATS_URL = os.getenv("NATS_URL", "nats://nats-server:4222")
|
||||
NATS_USER = os.getenv("NATS_USER", "admin")
|
||||
NATS_PASSWORD = os.getenv("NATS_PASSWORD", "package@321#")
|
||||
NATS_PASSWORD = os.getenv("NATS_PASSWORD", "")
|
||||
|
||||
# Domain Configuration (Env Vars from Deployment)
|
||||
NATS_STREAM = os.getenv("NATS_STREAM", "ORDERS")
|
||||
@@ -35,6 +36,10 @@ FILTER_SUBJECTS = os.getenv("FILTER_SUBJECT", "api.v1.mob.orders.*").split(",")
|
||||
|
||||
WORKER_CONCURRENCY = int(os.getenv("WORKER_CONCURRENCY", "10"))
|
||||
BASE_URL = os.getenv("EXTERNAL_BASE_URL", "https://jupiter.nearle.app")
|
||||
# Must stay comfortably above the outbound HTTP timeout below (30s), otherwise
|
||||
# JetStream can redeliver a message to another worker while this one is still
|
||||
# waiting on a slow-but-legitimate response, causing a duplicate forward.
|
||||
ACK_WAIT_SECONDS = int(os.getenv("ACK_WAIT_SECONDS", "60"))
|
||||
|
||||
# Endpoint mapping is still useful for constructing the target URL
|
||||
ENDPOINT_MAPPING = {
|
||||
@@ -56,6 +61,7 @@ js = None
|
||||
session: aiohttp.ClientSession = None
|
||||
running = True
|
||||
semaphore = None # Initialized in main
|
||||
active_tasks: set = set() # In-flight process_message() tasks, drained on shutdown
|
||||
|
||||
def signal_handler(sig, frame):
|
||||
global running
|
||||
@@ -80,7 +86,10 @@ async def forward_to_external(endpoint: str, payload: dict, api_key: str = None,
|
||||
if api_key:
|
||||
headers["Authorization"] = f"Bearer {api_key}"
|
||||
|
||||
data_to_forward = payload.get("data", payload)
|
||||
if "data" not in payload:
|
||||
print(f"?? Message for {endpoint} missing 'data' field, dropping.")
|
||||
return "DROP", None
|
||||
data_to_forward = payload["data"]
|
||||
|
||||
# Payload Normalization for logs
|
||||
if endpoint == "/live/api/v2/deliveries/createdeliverylog":
|
||||
@@ -166,7 +175,8 @@ async def process_message(msg):
|
||||
await msg.term()
|
||||
except Exception as e:
|
||||
print(f"? Critical Worker Error: {e}")
|
||||
await msg.nak()
|
||||
# Delay avoids a tight redelivery loop if this is a persistent bug
|
||||
await msg.nak(delay=5)
|
||||
finally:
|
||||
messages_in_flight.dec()
|
||||
duration = asyncio.get_event_loop().time() - start_time
|
||||
@@ -210,8 +220,9 @@ async def main():
|
||||
try:
|
||||
sub = await js.pull_subscribe(
|
||||
subject,
|
||||
durable=durable_name,
|
||||
stream=NATS_STREAM
|
||||
durable=durable_name,
|
||||
stream=NATS_STREAM,
|
||||
config=ConsumerConfig(ack_wait=ACK_WAIT_SECONDS)
|
||||
)
|
||||
subs.append(sub)
|
||||
except Exception as e:
|
||||
@@ -231,7 +242,9 @@ async def main():
|
||||
# Batch size 10, short timeout to keep loop responsive
|
||||
msgs = await sub.fetch(10, timeout=0.5)
|
||||
for m in msgs:
|
||||
asyncio.create_task(process_message(m))
|
||||
task = asyncio.create_task(process_message(m))
|
||||
active_tasks.add(task)
|
||||
task.add_done_callback(active_tasks.discard)
|
||||
except NatsTimeoutError:
|
||||
pass
|
||||
except Exception as e:
|
||||
@@ -244,6 +257,9 @@ async def main():
|
||||
except Exception as e:
|
||||
print(f"?? Fatal Error: {e}")
|
||||
finally:
|
||||
if active_tasks:
|
||||
print(f"?? Draining {len(active_tasks)} in-flight message(s) before shutdown...")
|
||||
await asyncio.wait(active_tasks, timeout=25)
|
||||
if session:
|
||||
await session.close()
|
||||
if nc:
|
||||
|
||||
Reference in New Issue
Block a user