Hi everyone, I just want to check my understanding of Big(O). Can someone check my thought process on the below?

Suppose you have a stream of integers you're operating on linearly. For every integer, you push the integer onto a heap. Such an algorithm's runtime complexity I believe would be represented as O(N*logN).

But what if, by design of the algorithm or of the problem constraints, you have a heap of fixed size? Suppose your heap is only designed to hold 3 elements. In other words, when you push a 4th element to the heap, it will also perform a pop.

In this circumstance, would the runtime complexity just be O(N)? I'm thinking that since the heap has a fixed size, it's basically a constant, aka O(N * logK) where K is the fixed size of the heap, but since K is a constant, you cut it out and make it O(N).