Well, "all' - it's a big word. Do we know in general how many metabolites have different organisms? But You definitely can identify and quantify many compounds. If You think about tens of compounds, You can use one of the most common technologies, if You think about hundreds - probably You will need to combine different techniques of metabolomics.
Regarding the specific part of pants, it should not be a problem. There are methods for leaf, seedling, shoot, berry, flower, seed etc metabolomics. Also, there are methods even for single cell metabolomics.
There is a fairly comprehensive book written outlining metabolomic methods and protocols - Plant metabolomics. Methods and protocols Nigel W. Hardy, Robert D. Hall. eds. 2012. Springer/Humana Press. £99 (hardback).
It gives an overview of the main analytical platforms: LC-MS, GC-MS, NMR and covers the advantages and limitations of each technique. Even with the full set of analytical platforms it would be difficult to identify and measure ALL of the metabolites present in any plant compartment.
the identification of metabolites by metabolomics is complicated and not without issues. Most platforms used for metabolomics now use accurate mass detection of ions, fragments of ions, adducts. By combinatorial computing, one can calculate the most probable net formula, in the sense of C6H12O2S2. By exploiting the naturral occurrence of isotopomers (formerly known as isotopes), one can eliminate most of the suggested compounds. But that leaves one with the fact that many compounds are so complex, that they contain several asymmetric carbons. And each of these asymmetric carbons introduces several configurations for the candidate molecule. What helps is a database with verified compounds, including sterochemistry. But even this is no guarantee for a correct identification. kind regards, Harrie Verhoeven.