Input SQS queue name not added to the URI properly when using AssumeRole

[jordo1138]

After locally fixing the input set region issue #249 , I found a interesting bug where when assuming the role, I am able to assume the role, successfully send listbucket request and then when it creates the sqs client and url to use, I get a 400 bad request because the queue name cannot be found for this wsdl.
But after inspecting using fluentd -vv it's clear the issue is that although my fluent.conf sets the SQS queue_name correctly, the SQS client POST request is to / instead of the /accountnumber/queue_name
The region for sqs HOST is also correct. See Image below. The log even shows the right queue name but for somereason didn't form the URL correctly. I have also properly granted my assumedRole access to the SQS queue in us-west-2 but no way to know if it's working since the POST request is incorrect.

The issue does not exist when using a key/secret key to another queue in the same region and the sqs client POST url is correctly formed, seems to be limited to assumeRole scenario
I believe the issue is due to the sqs queue not being in the same account as the assumed role. Is this a requirement? I have granted access to the assumed role to that sqs queue but seems it's a related issue.

[jordo1138]

modified to handle cross account id that owns the sqs queue when getting the Queue url per https://docs.aws.amazon.com/sdkforruby/api/Aws/SQS/Client.html#get_queue_url-instance_method
it's in the pull request now #249

[repeatedly]

Merged PR.