Daily account-usage refresh no longer stalls on large hosted accounts

Fixes a bug that stalled the daily monthly usage rollup on large hosted Tuist accounts. The job that refreshes each account’s cached monthly usage was building one giant multi-row INSERT that could blow past PostgreSQL’s 65,535 bind-parameter limit once the backlog of stale accounts crossed about 7,281, which raised a Postgrex error, dropped the whole batch, and let the backlog grow until the job exhausted all 20 retries. Worker inserts are now chunked into batches of 1,000 behind the scenes, so the daily refresh completes and accounts’ cached usage stays up to date again on big hosted instances. No behavior change for smaller accounts, which still fit in a single batch.