I see making a game engine as the illusion of progress on making a game. Making a game engine is fun and relatively easy. You have a check list of things to do. Each of them the solutions are relatively well known. so you do them and make tons of progress. You get a window open, then you get a triangle up, then you get a texture loaded, then you get some basic text for debugging, then you read the keyboard for input, etc etc. each day new stuff comes up and you think you’re making progress but really you haven’t even started making the game , you’re just reproducing what you could have already had if you’d started with an existing engine.

Then you start it hit the more tedious stuff. loading animated characters, blending animations on selective subtrees of a character hierarchy. Making a level editor. Adding quality of life feature to it like undo. Etc…

I’m not saying you shouldn’t do this. It’s fun to do. just don’t delude yourself that that’s making progress on your game. It’s instead making progress on a game engine. That’s a different thing.

I've shipped 18 games, 4 of them AAA. I wrote the engines for most fo them. I wouldn't do it again.

All that said, some nuance. If the game you are making is simple for some defintion of simple, Celeste, Dead Cells, Geometry Wars. Then making your own engine isn't much work and there maybe some benefits.

On the other hand, see all the tiles made with engines. Silksong is Unity. A Short Hike is Unity. Blue Prince is Unity. Valheim is Unity. Peak is Unity. Dredge is Unity. You don't need to make your own engine to make an indie game.