It seems like this issue is possibly still present even in v8, although it's in a worse state since there's no java client and zero visible indicators it's happening in the web UI - other than the scheduled tasks just not running at all.
In all honestly, this is TERRIBLE and it's more than enough alone to make us consider switching platform.
Lets face it, there's a million applications out there with schedulers built in that handle this scenario just fine, it appears to me that whatever issue is being described by iiliev is simply an architecture/design flaw on the part of VMWare. Sure it may explain why it's occurring, but it shouldn't occur full stop.
This application could be used for potentially critical and impacting jobs, so it needs to have a reliable way to schedule jobs long term. It should have been designed in a way that scheduled jobs could for example be run in a system context, or just some other way altogether that doesn't rely on expiring tokens that have a limit for renewal.