Knowing expected concentrations and matrix would help. Also, saying 'salt concentration' means something different in chemistry talk than common language. There are sodium salts, lithium salts, potassiium salts, etc. In addition, any anion could have been joined to any cation (sodium chloride, lithium chloride, potassium chloride, etc,. etc), so it would not be possible with a single analysis to know what the starting compound was, and then only with much experience with your process system - best guess would come from known contaminant into pure water matrix. 'Routine' means nothing more than what an instrument is standardized for, and is accompanied by accuracy and precision requirements, and assisted by whether it's on-line or run for only a few analyses at a time. Ion chromatography seems to be one way to get both the anion and cation concentrations easily. My minimum detactables by one IC are: Fl/Cl/SO4 0.14 ppt(rillion), and Na 0.14 ppt in a 3 ppm organic matrix. Minimum detectable Na by flame atomic emmission is 0.2 ppm. Graphite furnace is another option. If you're talking 'environmental levels' , AA and IC can do it, with much sample prep. But then, if it's that high, can you use a salinometer or conductivity meter?
The devices are complicated and fairly expensive, but entriely suitable for routine analysis.