Is this average/typical for a standard install?
It's the stats collection process (collectd). It runs every minute to collect the network, cpu, memory etc. stats and then generates the associated PNG images. It also has to aggregate up and roll over (tar/compress) the stats plus create/refresh the daily, weekly, yearly graphs at appropriate intervals.
It's likely to be the graph generation that is causing the CPU spikes, and these are being generated as static images even when they are not being viewed. It would be better (and entirely possible, although it would incur a small delay on low power systems) for the the graphs to be generated dynamically on demand as they are viewed by the user.
Is there a way to put the FN system a bit more to sleep? i.e. use less power?
Only by completely disabling all stats collection and associated image generation, which might be something you want to
request as an enhancement if this is of no value to you or you are monitoring stats externally. At the very least, dynamic rather than static image generation would be a worthwhile enhancement.
Dynamic graph generation may even need to become an essential requirement as more and more stats are collected and graphed, not to mention each volume and/or dataset results in an additional set of graphs etc. so as you add more volumes/datasets/interfaces/etc you increase the periodic load on the server while it grinds away creating graphs you very rarely - if ever - look at.
Finally, if you keep the the web GUI open in a browser tab frequent HTTP polls will keep the python process busy, and every HTTP access is logged to the file system. Even when no browser is connected, a watchdog process appears to be monitoring the state of the web server every 30 seconds or so (purpose, unknown as I very much doubt any "alarms" would go off should the web server become unavailable).
