fix: move sandbox ResourceQuota/LimitRange to gooseek-sandbox namespace
All checks were successful
Build and Deploy GooSeek / build-and-deploy (push) Successful in 3m26s
All checks were successful
Build and Deploy GooSeek / build-and-deploy (push) Successful in 3m26s
Kustomize was overriding namespace to gooseek for all resources including ResourceQuota and LimitRange meant for gooseek-sandbox. This caused CPU quota exceeded errors blocking all new pod creation (used 26.3/16 cores). Split sandbox-scoped resources into opensandbox-sandbox-ns.yaml applied directly, and deploy.sh now cleans up misplaced quota from gooseek ns. Made-with: Cursor
This commit is contained in:
@@ -82,6 +82,16 @@ if command -v envsubst &> /dev/null && [ -f "$ENV_FILE" ]; then
|
||||
envsubst < "$SCRIPT_DIR/monitoring.yaml" > "$SCRIPT_DIR/_generated_monitoring.yaml"
|
||||
fi
|
||||
|
||||
# Apply sandbox-namespace resources separately (not via kustomize to preserve namespace)
|
||||
echo ""
|
||||
echo "=== Applying sandbox namespace resources ==="
|
||||
kubectl apply -f "$SCRIPT_DIR/sandbox-namespace.yaml"
|
||||
kubectl apply -f "$SCRIPT_DIR/opensandbox-sandbox-ns.yaml"
|
||||
|
||||
# Clean up misplaced quota/limitrange from gooseek namespace (legacy fix)
|
||||
kubectl delete resourcequota sandbox-quota -n gooseek --ignore-not-found=true 2>/dev/null || true
|
||||
kubectl delete limitrange sandbox-limits -n gooseek --ignore-not-found=true 2>/dev/null || true
|
||||
|
||||
# Apply kustomization
|
||||
echo ""
|
||||
echo "=== Applying K8s manifests ==="
|
||||
|
||||
Reference in New Issue
Block a user