That is why I mean by "case by case", the secret thing that they don't tell you about game dev is that it is all about trade-offs and resources. It takes a lot of work (not just from me, remember everything needs to go through QA too) to make any sort of change ESPECIALLY if art is involved. So it really comes down to how valuable we feel a particular change will be vs the cost of making that change. Both games are different and have different problems we are trying to solve so "best fit" really comes down to a few things, and often that is what problems we are trying to solve.
In the Sona example, we are not seeing a problem with Sona being a powerful solo lane champion, so it would be a rather high-cost change with no clear goal in mind.
So generally an update needs to be solving some sort of problem, and the scale of that problem. Sometimes we will have a change in mind that is low cost but has a lot of potential upsides (Ahri update), and we are able to do it.
But even then we need to balance updates against other work that needs to be done (champion updates are done by the same people making new champions, and we don't want to push a champion's release back. While new items, runes, and balance changes, are done by a single group as well.
These priorities are handled by both designers and producers, and often include our insights teams (data analysis) to identify problems. These can also be adjusted as we go. This is a big reason why we don't talk about any changes with players too far in advance, things can always change and we don't want to mislead or give players the wrong impression.