Another slight mindset shift I am trying to make is to focus on problems, not solutions. This idea is parroted around a lot especially in online-entrepreneur-influencer circles, but I think it does have some value. As someone who can obsess over what the "best" solution to a problem is and make the entire process much more difficult for himself, this tweak may prove helpful. Instead of being occupied purely with the solution, pour more time and effort into learning about the problem. The classic questions of who, what, where, when, why, and how are all of the tools needed here.
The core hypothesis behind this idea is that a good solution should be evident from the problem, once you have enough data about the problem. Instead of languishing about large-scale architecture mistakes and refactors, focus on the specific issue at hand. If the original author is still on the team, reach out to them for some information. A quick slack message may illuminate something before unseen to you. Identify exactly what the issue is, and try to keep this definition targeted. Pinpoint exactly where in both the code and the product this issue occurs. Enumerate every place where this is the case, which may be helpful for testing at the end. Understand what decisions/mistakes led to this issue being checked-in to the code base. List out all of the different manners in which the issue can be caused. Is it reproducible? Is it deterministic? Given all of this information on top of the code itself, clearly explain how the issue occurs. From code decisions to what manifests to the end user, you should be able to explain each step along the way and what key pieces lead to the final result.
With this detailed of a view on the problem, a list of potential solutions should become immediately clear. Rank these solutions and get feedback from team members (if necessary) before moving forward with the best choice. Don't treat this is as a binding agreement: more information may become available while working on this fix that reveals another solution as a better choice. If unconstrained by time, feel free to switch to the new approach.
Instead of focusing on the solution, obsess over the problem. Gather info and make an educated decision on how to best tackle the problem. The end goal is to fix the issue, not to create an elegant solution. If the investigation into the problem is done right, this should happen largely on its own.
Well defined problems reveal elegant solutions.