A lot of modern cloud architecture is concerned with the business of making sure web requests don’t reach your servers if they don’t have to. Edge caching, content distribution networks, browser cache handling.
If you do it right, a high proportion of your site visits leave no trace in your logs that they were ever there.
Even if the CDN can't (for whatever reason) one could easily include a tracking pixel on every page that is marked as `Cache-Control: no-cache`, or insert a few lines of JS to do the same.
Presumably at some sort of additional cost, though. So then we’re into the business of weighing up whether to spend money on obtaining raw logs or purchasing the CDN’s own traffic analytics add on… or just going with a third party. This stuff isn’t just built in.
If you do it right, a high proportion of your site visits leave no trace in your logs that they were ever there.