- Add .AsSplitQuery() to ContentSyncHandler to prevent the cartesian product join caused by two collection .Include()s (Assets + AnimationBindings). EF was executing a single JOIN that multiplied result rows per card, causing replies to arrive 10+ seconds late and gateway timeouts. - Bump Gateway__RemoteDispatchTimeoutMs from 10000 → 15000ms in configmap to give worker pods headroom on cold starts (EF model compilation + first DB query can take 3-5 seconds on a fresh pod). NOTE: restart gateway pods after deploying to pick up the new timeout value: kubectl rollout restart deployment/sor-gateway -n stone-server Generated with [Claude Code](https://claude.ai/code) via [Happy](https://happy.engineering) Co-Authored-By: Claude <noreply@anthropic.com> Co-Authored-By: Happy <yesreply@happy.engineering>
26 lines
1.2 KiB
YAML
26 lines
1.2 KiB
YAML
apiVersion: v1
|
|
kind: ConfigMap
|
|
metadata:
|
|
name: sor-config
|
|
namespace: stone-server
|
|
data:
|
|
# ── Shared config (all pods) ─────────────────────────────────────────────────
|
|
# ENABLED_SERVICES and Services__Auth are NOT set here — each deployment sets
|
|
# them directly in its own env block so gateway and workers are independently
|
|
# configured without a shared ConfigMap override.
|
|
ASPNETCORE_URLS: "http://+:8080"
|
|
ASPNETCORE_ENVIRONMENT: "Production"
|
|
|
|
# WebSocket / gateway settings (consumed only by pods running the gateway service)
|
|
WebSocket__Path: "/ws"
|
|
WebSocket__MaxMessageSizeBytes: "1048576"
|
|
Gateway__AuthTimeoutMs: "10000"
|
|
Gateway__MaxMessagesPerSecond: "20"
|
|
Gateway__RemoteDispatchTimeoutMs: "15000"
|
|
|
|
# NATS message bus
|
|
Nats__Url: "nats://nats.stone-server.svc.cluster.local:4222"
|
|
|
|
# ── SOR.Admin ───────────────────────────────────────────────────────────────
|
|
# Admin inherits ASPNETCORE_URLS and ASPNETCORE_ENVIRONMENT from the same configmap.
|