返回 DeepSeek-Reasonix
migrate-drop-metric-users.sql
根目录 / workers / crash-report / migrate-drop-metric-users.sql
1 -- Apply: wrangler d1 execute reasonix-crash --remote --file=migrate-drop-metric-users.sql
2 --
3 -- Retires per-install metric detail. The 30-day COUNT(DISTINCT install_id) it
4 -- fed was the database's entire read bill: the primary key leads with `date`,
5 -- so the rollup's `date >= ... AND signal = ?` could only prune the window and
6 -- scanned all ~28M rows once per signal, eight signals an hour — ~5.4B rows
7 -- read a day for one dashboard module. The aggregate `metrics` tables keep the
8 -- same signals as event counts at ~180k rows total.
9 --
10 -- Deploy the worker first: it stops writing these tables and stops creating
11 -- them at runtime, so a DROP before that just gets rebuilt on the next request.
12 DROP TABLE IF EXISTS metric_user_rollup;
13 DROP TABLE IF EXISTS metric_user_rollup_state;
14 DROP TABLE IF EXISTS metric_users;
15 DROP TABLE IF EXISTS cli_metric_users;
16
16 lines SQL