Measure pools, queries, and failures

Monitor both Cloudflare and the origin database. Hyperdrive may reduce connection setup while a slow query or large pool still overloads the database.

Start with conservative pool settings. Measure active origin connections, query latency, timeouts, cache hit behavior, and error classes. Keep a direct recovery path and a way to disable query caching if stale results cause harm.

Run a bounded load test in a practice environment and correlate Worker observations with database connection and query metrics.

Measure one request with a narrow query and a request ID:

select id, title
from notes
where user_id = $1
order by created_at desc
limit 20;

Record application duration, database duration, rows returned, pool errors, and the query plan. Repeat after a warm request. If the result is slow, do not assume geographic distance is the cause: a scan, lock, or exhausted origin pool can dominate the same request.