Troubleshooting Request Connection

When configuring and running Omni Gateway, you might encounter connection issues.

To troubleshoot these issues, use the flexctl check http command. This command makes direct requests to both the outbound connection between the upstream service and Omni Gateway and the inbound connection between Omni Gateway and the client. Making direct requests isolates Omni Gateway connections, which allows you to localize the issue.

The following diagram demonstrates the difference between inbound and outbound requests.

Inbound connections are between the client and Omni Gateway and outbound connections are between Omni Gateway and the upstream service

Outbound Requests Between Omni Gateway and the Upstream Service

Isolating the connection between the upstream service and Omni Gateway enables you to test for connection between the two.

To make a direct outbound request to the upstream service:

Log into the Omni Gateway container or VM:

Linux:
```
ssh [-i sshkey.pem] <user>@<domain>
```
Docker:
```
docker exec -it <container-id> bash
```

Kubernetes:

kubectl -n <namespace> exec -it <pod-name> -- bash

Make a request to the upstream service:

flexctl check http -X=POST -d='{\"key\": \"value\"}' -H='Authorization:token' <http://upstream.domain>

Inbound Requests Between the Client and Omni Gateway

Isolating the connection between Omni Gateway and the client enables you to test if the connection issue is due to a load balancer or another system added between Omni Gateway and the client.

To make a direct inbound request to the client:

Log into the Omni Gateway container or VM:

Linux:
```
ssh [-i sshkey.pem] <user>@<domain>
```
Docker:
```
docker exec -it <container-id> bash
```

Kubernetes:

kubectl -n <namespace> exec -it <pod-name> -- bash

Make a request to Omni Gateway:

flexctl check http -X=POST -d='{\"key\": \"value\"}' -H='Authorization:token' http://localhost:<api-instance-port>/<path>