Well,
First thing that you should ask yourself:
Why I am drawing or want to draw a diagram about this?
Maybe you answer is :
To understand important/significiant communication between those elements.
Then we can say:
Oh maybe sequence diagram may help...
Buy may be you answer is:
Oh I am worry about topology of system components. How those things
will be deployed?
Then we can say that:
Oh, may be deployment diagram will help you...Draw one.
So depends on "context"...
UML diagrams will not solve your problems. They help you to understand tricky and hard part of your system and think about alternative solutions by visual modeling.
Modeling is not "self masturbating" activity. It is like "group sex". You will get most benefit when you do with others...
So the main problem is not to draw which diagram... The main problem is:
What kind of problem you have? With whom you wanto to find solution? What kind of benefit may you get from drawing a diagram?
For UML i definelty suggest Larman book: Applying UML and Patterns :Check at amazon
And lastly, if UML does not fit your purpose, be creative and pragmatic...Use textual desription or even create your "own" visual modeling :-) or do anthing which help you...