Thanks for the feedback @erik.hoffmann,
The areas you cited are ones that we are spending a great deal of time on refining. Creating a universal sync engine that works extremely well with all integrations and all users is pretty tough. Some integrations have proven to be quite challenging to work through their quirks and behavior and some users have some pretty demanding use cases. We also know that we currently have big gaps in our user feedback. We are working very hard (literally around the clock) on trying to provide the best possible experience for all integrations and our next major release will be bringing some great advancements in these areas.
1) Some integrations can be pretty restrictive in what they allow to be uploaded. OneDrive for Business, which I know you have a few accounts on, is probably the worst in this case. This means you can end up with files in "not allowed" that will never upload until a particular condition is corrected (illegal character or path length, for example). When a file is in a not allowed state, it can be difficult to notice if you are not looking for it.
Our badging on folders can cause some confusion, especially when there are items in both a synced and unsynced state. Generally speaking, if there are files that have yet to be uploaded in a structure, you should not see a checkmark on the file or the folders containing that file, but this requires taking a close look at the badging on items.
Our next major version is going to tweak the UI to make things more apparent and easily identifiable, especially if there are still items that have yet to sync or are having problems syncing. Additionally, if a particular service is having issues, it will be easer to identify so that a user can take corrective action.
2) Recursive sync should be working, in general, but this is another area where integration-specific exceptions can cause some frustrating behavior.
The gist is that the storage service is throwing errors that stop the recursive download process and require it to be kicked off again. Amazon Drive and OneDrive for Business tend to hit this behavior more than others, but all integrations can. The larger the amount of data/structure to be downloaded, the more likely you are to hit some sort of error that stops the download processing. Kicking off the recursive sync again will allow it to continue, but for some users it can error out again, quickly, and that can be pretty frustrating.
In this case, again, our next major version will improve this behavior significantly. It will better deal with these exceptions and be much more tolerant of transient issues, allowing the user's bulk download command to continue on, despite exceptions from the linked service.
3) I think this is similar to #2 and probably #1, where storage service exceptions can cause halting of sync processes. We are working very hard on improving odrive resilience to those exceptions. Additionally, general sync behavior will be more robust and predictable, despite integration quirks.
4) I think you may be referring to the warning messages you get about files not yet synced when trying to unsync a structure. That behavior can be triggered by the items in #1, #2, and #3 and basically boils down to issues trying to upload certain files that odrive isn't able to resolve on its own. As stated in #1, there are currently user feedback gaps when upload issues are encountered, and it ends up not being easy for the user to find, identify and then correct these. This will be a lot better in the next major version, as stated in the items above.
Everyone is different, so the degree to which these issues can be encountered will vary drastically depending on the integrations linked, the data you are working with, and the use cases involved. Lots of people use odrive without any issues, whereas others may hit some of these issues enough to cause some serious frustration, especially if using an integration that is on the more "quirky" side, like OD4B. This is why we offer the 30 day full refund on top of the 7 day trial, so that users can have a lot of time to see if odrive is the right fit for them.
We definitely understand all of this and we really want everyone's experience to be a great one, regardless of use case or link type. We are making great strides on combatting all of the above cases in our current development, which we are trying to get out as soon as possible.