Moving to the unofficial Answers, Ask the Community forum, so I can answer it, since it isn’t actually a programming problem.
#1, it is not a programming problem, it is a mechanical problem. Our team has it too. I noticed it last year with the BRAIN cpu as well.
The problem is that a servo wants to hold a specific position. When there is a spring and mass resonant load, the slightest movement displace the servo, which provides a restorative force, which stretches the spring, which oscillates the mass into an overshoot, which causes the oscillation you have noticed. To fix it mechanically, you need to change the resonant system:
The resonant system has components of Mass, spring, damper;
You can change the mass component by making the arm lighter/shorter.
You can change the spring component by making the arm more rigid/shorter.
You can change the damper component by adding some type of friction.
#2, Direct drive (connecting a clock hand) directly to the servo horn is a common idea, and a bad one when the load is massive, particularly this year when the servos have higher voltage from the cortex and burn out often. It is much better to use some Igus bearings, or something to hold the load, and then connect it to the servo with a linkage. Real bearings and a linkage would be a great idea for a gripper, probably not necessary for a clock pointer indicator.
Make sure you read the Q&A rules prohibiting “functional paint” for your clock face.