Exercise 3: rate() and a Mid-Window Counter Reset — Possible Solution
====================================================================

-- What rate() does when a counter resets mid-window --
--
-- rate() is specifically designed to detect a counter reset -- a
-- sudden drop in a counter's value between two consecutive samples,
-- which can only mean the process restarted and the counter began
-- counting again from zero, since ordinary counters are never allowed
-- to decrease any other way. When rate() detects this kind of drop
-- within its own window, it compensates for it automatically: instead
-- of treating the drop as a real negative change, it counts the
-- pre-reset portion and the post-reset portion as two separate
-- segments of genuine increase, and combines them into one correct,
-- still-positive average rate for the whole window. The service
-- restarting doesn't produce a nonsensical result -- it's handled as
-- a normal, expected event.

-- Why "current value minus value five minutes ago" would fail --
--
-- A naive manual calculation has no such reset-awareness built in --
-- it just subtracts the value from five minutes ago from the current
-- value, with no concept of "this counter might have been reset in
-- between." If the counter was, say, at 50,000 five minutes ago,
-- dropped to 0 on restart partway through the window, and has since
-- climbed back up to 8,000, that naive subtraction computes
-- 8,000 - 50,000 = -42,000 -- a clearly wrong, negative result for a
-- value (request count) that can never actually decrease. This is
-- exactly the failure case the chapter names rate()'s own
-- reset-handling as solving: rate() would instead correctly report a
-- positive rate reflecting the real traffic that occurred both before
-- and after the restart, while the manual subtraction produces a
-- meaningless negative number with no correct interpretation at all.

WHY THIS WORKS AS AN ANSWER
------------------------------
This explains rate()'s own reset-detection mechanism concretely
(treating pre- and post-reset segments as separate genuine increases),
then works through a specific numeric example showing exactly how a
naive subtraction produces an impossible negative result, directly
demonstrating why rate()'s reset-awareness genuinely matters rather
than just being a nice-to-have detail.