✨ New: Try our AI‑powered Search (Ctrl + K) — Read more

Complete Port Matrix and Firewall Rules

Prev Next

This article is intended to assist Customer Experience Engineers, Network Engineers, Field Engineers, with complete port requirements for firewall configuration and network planning for the Crosswork Assurance platform. Use this guide when configuring firewalls, security groups, or troubleshooting connectivity issues. Getting ports wrong is one of the most common causes of installation failures and connectivity problems.

Network Topology Diagram

This diagram shows the network architecture and how different components communicate. Use it to understand which ports need to be open between which components:

Quick Reference

If you need a quick overview, this table summarizes the port categories. Detailed tables follow for each category:

Category Ports Direction
External Access 443, 2443, 3443, 4443, 30000 Inbound
Inter-Node 2380, 4789, 6443, 9091, 9443, 10249-10256 Bidirectional
Local Processes 2380, 4789, 6443, 9091, 9443, 10249, 10250, 10256 Local only

External Access Ports

These ports must be accessible from outside the cluster for user and administrative access. The specific ports depend on whether you're using DNS mode or No-DNS mode.

DNS Mode (Recommended)

In DNS mode, all applications share port 443 and are differentiated by hostname. This is the recommended configuration for production deployments:

Port Protocol Application Description
443 TCP Tenant UI Main application access (https://{domain})
443 TCP Deployment UI Deployment dashboard (https://deploy.{domain})
443 TCP Zitadel (auth) Identity management (https://auth.{domain})
443 TCP Grafana Metrics dashboards (https://grafana.{domain})

Note: In DNS mode, all applications share port 443 and are differentiated by hostname.

No-DNS Mode

In No-DNS mode, each application uses a different port. Use this when you can't configure DNS records:

Port Protocol Application Access URL
443 TCP Tenant UI https://{external-ip}
2443 TCP Deployment UI https://{external-ip}:2443
3443 TCP Zitadel (auth) https://{external-ip}:3443
4443 TCP Grafana https://{external-ip}:4443

Admin Console

The Admin Console is always accessible on port 30000, regardless of DNS mode:

Port Protocol Application Description
30000 TCP KOTS Admin Console Installation management (https://{ip}:30000)

Sensor Ports

These ports are used for sensor communication. Sensors send telemetry data to PCA and receive configuration updates:

Port Protocol Service Description
443 TCP RoadRunner Telemetry ingestion (gRPC over TLS)
443 TCP Sensor-Orchestrator Configuration distribution

Inter-Node Communication

These ports must be open between all cluster nodes. They're used for Kubernetes control plane communication, etcd clustering, and pod networking:

Port Protocol Component Description
2380 TCP etcd etcd peer communication
4789 UDP VXLAN Pod network overlay
6443 TCP Kubernetes API API server
9091 TCP kURL kURL proxy
9443 TCP kURL kURL proxy (TLS)
10250 TCP Kubelet Kubelet API
10249 TCP kube-proxy Metrics
10256 TCP kube-proxy Health check

Local Process Ports

These ports are used for local communication only and don't need to be opened in firewalls between nodes:

Port Protocol Component Description
2379 TCP etcd etcd client connections
7443 TCP kURL Registry proxy
9099 TCP Calico Felix health
10248 TCP Kubelet Health check
10257 TCP kube-controller-manager Secure port
10259 TCP kube-scheduler Secure port

Internal Service Ports

These are the ports used by PCA services for internal communication. You generally don't need to configure these in external firewalls, but they're useful for troubleshooting:

Service Port Health Endpoint Expected Response
Gather 80 /health 200 OK
Fedex 80 /health 200 OK
Zitadel 8080 /healthz 200 OK
Druid Coordinator 8081 /status 200 OK
Health Clinic 80 /health 200 OK

Firewall Rule Examples

Use these examples as templates for configuring your firewall. Adjust IP addresses and CIDRs to match your environment.

iptables (Linux)

# External access (DNS mode)
iptables -A INPUT -p tcp --dport 443 -j ACCEPT
iptables -A INPUT -p tcp --dport 30000 -j ACCEPT

# Inter-node communication (replace with your node IPs)
NODE_IPS="<node-1-ip> <node-2-ip> <node-3-ip>"
for NODE_IP in $NODE_IPS; do
  iptables -A INPUT -s $NODE_IP -p tcp --dport 2380 -j ACCEPT
  iptables -A INPUT -s $NODE_IP -p udp --dport 4789 -j ACCEPT
  iptables -A INPUT -s $NODE_IP -p tcp --dport 6443 -j ACCEPT
  iptables -A INPUT -s $NODE_IP -p tcp --dport 10250 -j ACCEPT
done

firewalld (RHEL/CentOS)

# External access
firewall-cmd --permanent --add-port=443/tcp
firewall-cmd --permanent --add-port=30000/tcp

# Inter-node (create zone for cluster)
firewall-cmd --permanent --new-zone=pca-cluster
firewall-cmd --permanent --zone=pca-cluster --add-source=<node-cidr>
firewall-cmd --permanent --zone=pca-cluster --add-port=2380/tcp
firewall-cmd --permanent --zone=pca-cluster --add-port=4789/udp
firewall-cmd --permanent --zone=pca-cluster --add-port=6443/tcp
firewall-cmd --permanent --zone=pca-cluster --add-port=10250/tcp

firewall-cmd --reload

Cloud Security Groups (AWS/GCP/Azure)

Use these rules as a template for cloud security groups:

# Inbound Rules
Type        Protocol  Port    Source
HTTPS       TCP       443     0.0.0.0/0 (or restricted CIDR)
Custom TCP  TCP       30000   Admin IP range
Custom TCP  TCP       2380    Cluster security group (self)
Custom UDP  UDP       4789    Cluster security group (self)
Custom TCP  TCP       6443    Cluster security group (self)
Custom TCP  TCP       10250   Cluster security group (self)

Troubleshooting Network Issues

Use these commands when you suspect network connectivity problems.

Verify Port Connectivity

# From outside cluster
nc -zv <IP_ADDRESS> 443
nc -zv <IP_ADDRESS> 30000

# On a node to check listening ports
ss -tlnp | grep -E "(443|2443|3443|30000|6443)"

# Check inter-node connectivity
nc -zv <other-node-ip> 6443
nc -zv <other-node-ip> 2380

Common Issues

This table maps common symptoms to their likely network causes:

Symptom Check
Can't access UI Port 443 open, DNS resolution, nginx pod running
Can't access Admin Console Port 30000 open, kotsadm pod running
Nodes can't join cluster Ports 6443, 2380, 10250 open between nodes
Pod networking issues Port 4789/UDP open for VXLAN overlay

© 2026 Cisco and/or its affiliates. All rights reserved.

For more information about trademarks, please visit:
Cisco trademarks 
For more information about legal terms, please visit:
Cisco legal terms