Index the event log's resubmit count with deleted_at (closes #325)
check / check (push) Successful in 3m14s

The resubmit count matches resubmitted_from_id against a page of event
ids, and GORM adds deleted_at IS NULL. With no statistics, SQLite
preferred the deleted_at index once the page held five or more ids, so
every event log page read every live event. deleted_at now joins
resubmitted_from_id in idx_events_resubmitted_from_id, as for the other
event-tier indexes, and the plan test checks the statement with a full
page of ids.

Model: opus-5-5
This commit is contained in:
2026-10-02 19:10:20 +00:00
parent 35d2f28c67
commit ded9b0398e
3 changed files with 46 additions and 7 deletions
+7 -4
View File
@@ -1878,16 +1878,19 @@ tags, so `AutoMigrate` creates them on a fresh database:
| `deliveries` | `event_id`, `deleted_at` | The event log, which loads each event's deliveries, and retention, which counts and deletes the deliveries of expired events |
| `delivery_results` | `delivery_id`, `deleted_at` | The event log, which loads the attempts of a page's deliveries, and retention, which deletes the attempts of expired events |
| `events` | `deleted_at`, `created_at` | The webhook page's statistics, which count recent events |
| `events` | `resubmitted_from_id`, `deleted_at` | The event log, which counts the events resubmitted from each event on a page |
| `events` | `created_at` | Retention, which selects expired events by age |
GORM's soft delete adds `deleted_at IS NULL` to these queries; retention
leaves it out. SQLite keeps no statistics on these tables, and without them it
rates the `deleted_at` index, which every live row matches, above an index on
a column matched against several values or compared with a range. So every
index but the last also covers `deleted_at`. It comes second, so that
retention can use the index without it, except in `events`, where the
statistics compare `created_at` with a range (`>=`) and SQLite narrows by a
range only on the last column it uses.
index but the last also covers `deleted_at`. It comes second in the `event_id`
and `delivery_id` indexes, so that retention can use them without it. The event
log's count, the one query on the `resubmitted_from_id` index, always carries
`deleted_at IS NULL` and uses both columns. In the statistics' `events` index
`deleted_at` comes first, because they compare `created_at` with a range (`>=`)
and SQLite narrows by a range only on the last column it uses.
#### Common Fields