For example I deal with volume. In some parts tracks sits nice with other sounds and etc, e.g. works. In some parts of the track when some instruments drop then some tracks sounds too loud so I think maybe I need to automate volume on other tracks.
In most real-life mixes - studio recordings and live sound as well - different parts of a song (or a concert) require changes to channel/track balances throughout the song (or concert), so you can't just set all the faders and other settings to some static balance and expect it to work throughout the song (concert). It's the job of a mixing engineer to listen to the mix and make adjustments. I suppose this is most commonly needed for lead vocals, which have to be loud enough to be heard clearly at all times.
In modern DAW studio work, adjusting track balance during the song is usually done with automation, either by recording fader movements ("riding" the faders together with the performance) or drawing curves with the mouse or other means. But in old style analog recording and in live audio, moving and riding the faders is done live by a mixing engineer, reacting to changes as they happen. (Except in some big productions there may be automation even live, but then the show must be locked/synced to a pre-made sequence.)
So: a mix is not static.
In theory, a group of good performers is able to self-adjust their balance, if they can hear each other properly and if they know that there's no-one at the mixing desk. Try it: set all faders to some static position and then sing and play your tracks so that it doesn't need mixing. In practice this requires quite a lot of experience.