:warning: This concerns the synchronization of AMPF enabled Projects and Storages.

### What does the synchronization job does:

A **lot** of stuff, namely:

1. Creates Project Folder if it isn't there Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. Renames the folder if the need be Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. Removes access to the folder for all archived projects Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 4. Ensure that OpenProject admins, that have accessed the storage, have access to all folders Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 5. Remove permissions from users that lost access to that project Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 6. Add permissions to users that have been added to the project Begin of the insertion


End of the insertion Begin of the deletion

End of the deletion All this is run for every `Storage` with `automatically_managed_folders_enabled` defined on the following occasions:

1. Every 60 mins (CronJob) Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. On Member Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 1. Create Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. Update Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. Destroyed (poor member) Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. On Project: Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 1. Updated Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. Renamed Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. Archived Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 4. Unarchived Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 4. On OAuthClientToken Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 1. Created Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 5. On Role: Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 1. Update Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. Destroyed Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 6. On storage: Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 1. Created Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. Updated Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. Destroyed Begin of the insertion


End of the insertion Begin of the deletion

End of the deletion We have some debouncing going on to deal with this, but still it is a lot of triggers for a fairly expensive jobs.

### Proposal

We can first, break down the monolithic service into its parts, for example:

1. `CreateManagedFolderService` Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 2. `RenameManagedFolderService` Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 3. `HideManagedFolderService` Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 4. `AddUserToManagedFolderService` Begin of the insertion

End of the insertion Begin of the deletion
End of the deletion 5. `RemoveUserFromManagedFolderService` Begin of the insertion


End of the insertion Begin of the deletion

End of the deletion The current implementation can then make use of those for its operations (maybe?)

Create new jobs for each of these and hook those to the events.

Once they are into place we can reduce the long expensive one to, let's say once every couple of hours or less. Begin of the insertion

DONE ALREADY:End of the insertion Begin of the deletion

End of the deletion Secondly we can break down the job on per storage basis:
1\. One job == One storage Begin of the insertionsynchronization. End of the insertion Begin of the deletion synchronization.End of the deletion
2\. Then on top of this job we have one to synchronize all storages.
3\. On each event we schedule synchronization of only involved storages. Begin of the insertion

### Next steps

Review it and identify next simplification steps.End of the insertion