By this I mean these types of scenarios:
The instance goes down (for example after an update). Will workflows mid execution carry on from where they went down?
A node fails (a remote service does not respond). Will the node try again until it succeeds?
The workflow contains a long running interval (e.g. after 10 days if this value is false send this email). Is this reliable?