Performance Monitor and Profiler are both graphical tools that I wouldn't recommend ever running directly against production machines. Instead, create Data Collector Sets and Server-side Trace from each of the tools respectively. These are command line equivalents of the tools that you can use to output to a file. That allows you to collect the data without running the GUI on the servers. You can then use the GUI software on a different machine to browse and aggregate the data. And, if you're on 2008 or better, I'd strongly recommend extended events over trace/profiler. The events are more lightweight and the filtering mechanisms occur at data collection time, not after collecting the data. Makes a huge difference in the impact on a system. They're much safer.
Can you elaborate on your question? Performance Monitor is very light weight, running a profiler trace against production is ok in many situations depending on the load on the server and what all you are collecting. It is a primary tool for many of us for collection workloads to do analysis on. Extended events are also good for such data. It is highly recommended though to run the trace on another station other than the actual server.