Add exception details to the toComposeNetwork log.

[shencangsheng]

Description

When creating a compose, the latest version still frequently encounters issues where containers cannot join the network. #11601

Issue

I found in the source code that when there is an exception during network allocation, an empty network is assigned directly without logging.

Proposal

I would like to add a log message in this case to help identify and resolve the issue.

[ndeloof]

IMHO would be legitimate to just return err
As inspected container which we guess network from exists, not being able to inspect network should just demonstrate and API call or engine failure

[shencangsheng]

@ndeloof

As described, the containers orchestrated by docker compose up in my production environment have experienced network loss issues more than once. This problem persists even in the latest version and has troubled me for a long time. I have not been able to find any relevant error logs anywhere to pinpoint the issue. Even in the cli, no anomalies are recorded, which is very frustrating for me.

Would it be more appropriate to print the logs in the NetworkInspect method of cli?

// NetworkInspect returns the information for a specific network configured in the docker host.
func (cli *Client) NetworkInspect(ctx context.Context, networkID string, options network.InspectOptions) (network.Inspect, error) {
	networkResource, _, err := cli.NetworkInspectWithRaw(ctx, networkID, options)
	if err != nil {
		logrus.Warnf("Failed to inspect network %v", err)
	}
	return networkResource, err
}

[ndeloof]

generate command is not designed to be used as a diagnostic tool! You can wrap engine error to give more context, but I don't expect this will give you more than "not found":

  return fmt.Errorf("failed to inspect network %s: %w", networkID, err)
``