Caprock Files

Changed

• The main screen's heaviest request went from 0.61s to 0.001s. /v1/history answers "everything, ever" — four aggregates over the whole events table — and five components on that one screen ask for it on their own timers: the lifetime strip, the breakdown panel, the share card, the share nudge, and the screen itself. A single open tab produced bursts of identical requests and each one was computed from scratch.

Requests for the same range that arrive while one is already running now wait for it and share its result, and a settled result is reused for three seconds — shorter than the fastest poller, so nothing on screen is visibly behind. Errors are never cached, and a caller hanging up no longer takes the answer away from the callers waiting behind it.

Two smaller fixes found by the same measurements: the session count is grouped rather than COUNT(DISTINCT), which runs off a covering index instead of sorting every row into a temp B-tree (0.14s → 0.02s at range=all, no slower at any other range); and idle database connections are kept rather than reaped after five minutes. Each connection carries its own page cache, and a fresh one is cold — the same query measured 378ms on a connection's first use and 26ms on its second.