reduce reliance on 'team refresh'

[bobheadxi]

I implemented #484 as a part of team refresh, but @cheukyin699 brought up some good points against it:

Okay so here's the thing. /rocket team refresh takes around 20 seconds to execute (last I checked). The more members we have in Launch Pad, the longer it takes, O(n). We'll still need ways for users to receive the permission sharing automatically, whether if it is a user just adding their emails, or a team adding a folder ID into the system.

I am very adamant about not using /rocket team refresh unless we really need to because ... there are lots of side effects that I don't know about, that may be disruptive. For example, today I found out that it would forcefully add a user back into the organization if the user was in ddb but was already out of the org (one possibility would be if the user ID was still in a team in ddb).

I wasn't aware of the difficulties here because I didn't delve very deeply, but this can certainly get quite concerning, and at the moment I'm not actually 100% sure #484 will work very well once rolled out to our ~200 person Rocket deployment. Will keep an eye on this, but I'm guessing this task will encapsulate:

• decouple all team sync from refresh
• decouple drive permissions sync from refresh => Trigger Google Drive shared folder sync on user email edit #498

[bobheadxi]

From #504

I know we talked about possibly dropping refresh (#490), but we still need a way to perform the initial sync - I imagine refresh might stay around purely for use cases like this.

Updated title accordingly