It is always possible to create a scenario in which secondary becomes a lot behind primary when using synchronous commit thus resulting in high estimated recovery time even when taking log backups every minutes. This happens when primary is generating log at a higher pace then secondary can apply the log at. We have been able to reproduce the scenario each time with the same test in which estimated (and actual) recovery time goes on increasing to as high a value as one wants.
Operating System Language
Steps to Reproduce