Overhead for using iostat, top, and iperf commands

I'm currently looking at the best ways to monitor and report system statistics. So far, the best plan seems to grab output from iostat, top, and iperf, roll all the data up into a json object, and send it to a web server (doing this every second or so to deliver statistics in real-time).

Will the overhead from the three commands + json upload to a server + the program itself significantly impact the system and monitoring statistics? I'll use your guys' judgment as to the definition of "significant"; I'm new to system monitoring.

To go a little bit off path, I'm also curious as to whether this is a viable design for analyzing the performance of a system and/or applications on a system. If appropriate, I can also ask a new question; it's tough starting out in anything because the 'what to ask' is just as tough to figure out as the answers themselves.

- - , .
.

Recent Questions...