Probably Im just stupid but I want to understand how to make thread-safe my operations with db.
For example, I have next code

public async Task<Result> ProcessObjectAsync(long id)
{
  SomeObject obj = await _objProvider.GetObjectAsync(id); // gets obj from MongoDB
  if (obj.FieldOne == "foo")
  {
    obj.FieldTwo += 100;
    obj.FieldOne = obj.FieldTwo >= 1000 ? "stop" : "bar";
  }
  else if (obj.FieldOne == "bar")
  {
    obj.FieldThree += 222;
    obj.FieldOne = "foo"; 
  }
  Result res = new()
  {
    Name = obj.FieldOne,
    Sum = obj.FieldTwo + obj.FieldThree
  };
  await _objProvider.SaveObjectAsync(obj); // saves modified object to MongoDB
  return res;
}

_objProvider just some class that simply gets and sets objects in MongoDB
I need to make this code thread-safe, because now it is possible that two requests received in similar time can execute code in one block two times because updated object still not in the db

I see examples with transactions but in all examples they show just very basic CRUD things like “updating object with new data without using current state (fetching it)” but I need to know state of object, so firstly I need to fetch it, do some modifications and after that save it. I need to write some “lock system” around it or I just doesn’t understand how I can use database mechanisms for thread-safety even for those non-basing operations

I will be great if you provide me some examples how to do such things atomically or thread-safe