Total tokens count input and output only; cache tokens are shown separately

The headline “total tokens” figure — company-wide and per provider — sums input and output tokens. Cache-write and cache-read tokens are excluded from it, but stored and shown separately: in the breakdown by token type, in the per-model table and as their own chart views.

Why: agentic coding re-reads large repository contexts thousands of times, so cache reads outnumber input and output by one to two orders of magnitude. Summed into one figure they would push the total into an extreme range, flatten real fluctuations into noise and make the figure incomparable over time. Excluding them keeps the total meaningful; keeping them visible elsewhere means nothing is dropped.