Port 5228: Understanding Mobile Messaging Connectivity

One managed phone has a working service connection but does not display a particular notification. Another cannot maintain the connection at all. These cases require different investigations: TCP traffic to port number 5228 may be part of the service path, while message creation, device policy, and application behavior determine what happens afterward.

Port 5228 is associated with outbound connections used by supported Google mobile messaging services, including Firebase Cloud Messaging (FCM) on Android. It is a destination port for a client connection, not a reason to expose an inbound service on every device. For connections to TCP port number 5228, the network team should verify the service’s requirements and its own policy before making any change.

The connection is one part of the path

An application server submits a message to the messaging service; the service then attempts delivery to the intended device and application. The device needs suitable connectivity to maintain or establish the service connection. A successful TCP session demonstrates only that a network connection was established to the tested destination. It does not prove that the correct message was sent, that the target registration is valid, or that the application displayed it.

WSO2’s Android documentation, published in 2018, describes FCM as typically using TCP 5228 and sometimes 5229 or 5230. Other service traffic can use HTTPS on TCP 443; the exact role and fallback behavior depend on the service and client. A connection on 443 should not be assumed to substitute for every required path. Check the provider’s documentation for the deployed products and the organization’s approved network design.

The port number alone also does not identify a trustworthy destination. Review the expected service endpoints using vendor guidance, applicable device-management documentation, and network policy. Destination addresses and service dependencies can change, so an old list copied into a rule may become incomplete.

Trace the managed device’s network

For a device connecting on 5228 over TCP, record the affected device, time, network, VPN state, and whether other devices using the same application are affected. Compare an approved corporate Wi-Fi path with another authorized path only when organizational policy permits. A difference can help locate the issue, but it does not by itself prove that a firewall is responsible.

Review DNS resolution, routing, VPN policy, proxy behavior, and firewall logs at the relevant points. Some policies treat device-generated traffic and application traffic differently. Stateful devices and network address translation can also affect long-lived connections if their timeouts are inappropriate for the service. A review of TCP 5228 connectivity should include the return traffic and the behavior of intervening network devices, not only a single outbound rule.

Changes to network controls belong in the normal approval process. The responsible team can assess the required destinations, protocols, and related ports, then implement and test the smallest policy consistent with vendor guidance and organizational requirements. Opening TCP ports broadly is not equivalent to diagnosing a service dependency.

Separate service connection from notification delivery

If the connection is available, follow an approved test message through the sending application, service response, device, and receiving application. Check the correct device registration, application configuration, and message type. A message received by the device may be handled differently depending on whether the application is active or in the background. Notification permission, device-management settings, battery behavior, and application code can affect presentation without indicating a blocked connection.

Collect only the logs and identifiers needed for the investigation, and protect them according to policy. Do not request credentials or change a user’s notification restrictions to force a test. The objective is to identify where the expected event stops, then involve the owner of that component.

Diagnose in sequence

First confirm the expected message and affected device, then document the network and managed policy in effect. Check the service connection and relevant logs for the approved destinations, including traffic using TCP destination 5228 and other documented paths. If connectivity succeeds, trace the test message through registration, service handling, device receipt, and application presentation. Record timestamps and results at each step, make only approved corrections, and repeat the test. Port 5228 diagnostics are complete when both the connection and the intended delivery have been checked independently.