Apache web server logs aren't retained for IIIF, but have been historically used to gather usage information.

http://staff.lib.virginia.edu/systems/libsys/reqd.php?id=25965&rt=2&e=1

We don't want to retain all request data, but we do want to:

• measure relative use (omit bots or monitoring if possible)
• rank resources by popularity (omit bots or monitoring if possible)
• determine (and rank) where accesses come from
• more TBD