The aws-swf component does not set a default "Execution Start To Close Timeout" or "Task Start To Close Timeout" when workflows are newly registered.
This has the effect of scheduled workflows immediately moving to the TIMEOUT state (and being tracked as timed out in the AWS SWF UI).
For newcomers to the aws-swf component this is particularly confusing. My work-around was to explicitly create the workflow type in the AWS SWF UI before camel had a chance to register it with proper timeouts (the UI requires you to set a timeout, which may mean there's also a bug in the AWS API). Another approach looks to be related to the workflowTypeRegistrationOptions uri option on the aws-swf endpoint, though I haven't validated that approach as a work-around for this issue, so explore at your own risk.