What causes EndpointNotFoundException?

A firewall rule on the host VM or network perimeter is blocking the net.tcp port 8060 used for inter-node communication. The hostname or fully qualified domain name (FQDN) of the peer node is not resolvable from the affected node's DNS configuration. The DIAHostService or Self-hosted Integration Runtime Host service on the target node is stopped or crashed. Network Security Group (NSG) rules or corporate proxy settings are intercepting or dropping TCP traffic on port 8060 between SHIR nodes

How do I fix EndpointNotFoundException?

Step 1: On the Inactive SHIR node's VM, open Event Viewer and navigate to Applications and Services Logs > Integration Runtime. Filter for error logs and confirm the presence of the System.ServiceModel.EndpointNotFoundException referencing net.tcp:// :8060/ExternalService.svc/.. Step 2: Verify that port 8060 is open between all SHIR nodes by running 'Test-NetConnection -ComputerName -Port 8060' from PowerShell on the affected node. If the connection fails, update inbound and outbound firewall rules on both VMs to allow TCP 8060.. Step 3: Confirm that the peer node's hostname or FQDN resolves correctly from the affected node using 'Resolve-DnsName '. If resolution fails, add a hosts file entry or fix DNS configuration in the virtual network.. Step 4: On the peer (target) node, verify that the 'Self-hosted Integration Runtime Host' and 'DIAHostService' Windows services are running. If stopped, start them and check the event log on that node for its own errors.. Step 5: If nodes reside in different subnets or VNets, review Network Security Group rules to ensure TCP port 8060 traffic is explicitly allowed between the SHIR node IP ranges, then re-check node status in the ADF portal.

High severityconnectivityAzure Data Factory →

Power BI Refresh Error:
EndpointNotFoundException

What does this error mean?

A Self-Hosted Integration Runtime node reports an EndpointNotFoundException because it cannot establish a net.tcp connection to another node in the same IR cluster, causing the node to appear Inactive. This indicates a network communication failure between SHIR nodes.

Common causes

1A firewall rule on the host VM or network perimeter is blocking the net.tcp port 8060 used for inter-node communication
2The hostname or fully qualified domain name (FQDN) of the peer node is not resolvable from the affected node's DNS configuration
3The DIAHostService or Self-hosted Integration Runtime Host service on the target node is stopped or crashed
4Network Security Group (NSG) rules or corporate proxy settings are intercepting or dropping TCP traffic on port 8060 between SHIR nodes

How to fix it

1Step 1: On the Inactive SHIR node's VM, open Event Viewer and navigate to Applications and Services Logs > Integration Runtime. Filter for error logs and confirm the presence of the System.ServiceModel.EndpointNotFoundException referencing net.tcp://<hostname>:8060/ExternalService.svc/.
2Step 2: Verify that port 8060 is open between all SHIR nodes by running 'Test-NetConnection -ComputerName <peer-hostname> -Port 8060' from PowerShell on the affected node. If the connection fails, update inbound and outbound firewall rules on both VMs to allow TCP 8060.
3Step 3: Confirm that the peer node's hostname or FQDN resolves correctly from the affected node using 'Resolve-DnsName <peer-hostname>'. If resolution fails, add a hosts file entry or fix DNS configuration in the virtual network.
4Step 4: On the peer (target) node, verify that the 'Self-hosted Integration Runtime Host' and 'DIAHostService' Windows services are running. If stopped, start them and check the event log on that node for its own errors.
5Step 5: If nodes reside in different subnets or VNets, review Network Security Group rules to ensure TCP port 8060 traffic is explicitly allowed between the SHIR node IP ranges, then re-check node status in the ADF portal.

Frequently asked questions

Do all nodes in a self-hosted IR cluster need to communicate on port 8060?

Yes. SHIR nodes in a multi-node cluster use net.tcp on port 8060 for inter-node coordination and health signaling. If any node cannot reach another on this port, the unreachable node will appear as Inactive in the ADF portal regardless of its own service health.

The peer hostname in the error message looks correct — why would DNS resolution still fail?

In Azure VNets, DNS resolution between VMs can fail if custom DNS servers are configured on the VNet or subnet and those servers do not have records for the SHIR host VMs. Verify that all SHIR node VMs are registered with the correct DNS zone, or add explicit entries to each node's hosts file as a workaround.