Acknowledgements are mostly yours to do with as you please---in my experience in computer science, at least, there are few hard and fast rules.
For my own part, I typically keep my acknowledgements relatively spare, reserving them for people who made significant and meaningful contributions, but who did not rise to the level of authorship. As such, I tend to not acknowledge anonymous reviewers, people who gave feedback that didn't result in large changes in the manuscript, etc. You should feel free to be more liberal in your acknowledgements if you choose, however---the only real limitation is if either a) your manuscript may have length limitations or b) you do something that can be taken as silly (which can be OK for a thesis but generally not so much in a peer-reviewed manuscript).
It is pretty much always the case, however, that acknowledgements say what a person is being acknowledged for. If you say that, then people will know how to calibrate the contribution. If you have a hard time figuring out what to write, that may be a sign that the person didn't do enough to be acknowledged.