[Tracking] Migration to task scheduler

[AlCalzone]

As discussed on Slack, there are issues with SmartStart inclusion when too many things are happening at once. For example multiple interviews, plus automatic inclusion of nodes. It seems that the active communication with nodes is interfering with the passive communication that's happening under the hood.

This is somewhat parallel to #903 and #884, but we could avoid this by grouping communication in batches (some of which can be paused) and scheduling them. We currently only queue individual transactions, leading to seemingly chaotic queues during busy interview times. By scheduling larger batches, interviews of individual nodes wouldn't be interrupted by others.
User-initiated communication would not be affected.

Use cases:

• Inclusion and security bootstrapping (not pausable, other than the times where user interaction is happening)
• Exclusion
• Node interview (pausable for high-priority batches)
• Network healing feat: rework route rebuilding to use task scheduler #7196
• Health checks
• OTA updates refactor: migrate OTA firmware updates to tasks #7256
• dropping everything for switching to bootloader fix: remove all tasks when hard-resetting or entering bootloader #7255
• dropping everything for hard-resets fix: remove all tasks when hard-resetting or entering bootloader #7255