* op-supervisor: Introduce head tracker

* op-supervisor: Move log db to subpackage of db.

* op-supervisor: Route all updates through a common db that can track heads

* op-supervisor: Remove unused error.

* op-supervisor: Remove operations - it fits into a later PR.

* op-supervisor: Fix semgrep

* op-supervisor: Move resuming databases into ChainsDB so it can later update the chain heads too.