Prober: claim only candidates of settled releases, without the join
Test and deploy / test (push) Successful in 48s
Test and deploy / edge (push) Successful in 1m59s
Test and deploy / deploy (push) Failing after 4m2s

The settler marks a settled release's candidates ready (and unreadable when
it had no readable name). Claims took the candidates by joining releases
from random points of the queue; in its newer half nothing had settled, so
a slice scanned to the end (0.3 s each, 20 slices and 5 tiers a claim) and
the prober stood still: 2.4M waiting, 0 probed a minute. Now an index over
the claimable rows only, the slices drawn within them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Daan MeijerandClaude Opus 5.5 committed 2026-10-12 12:12:59 +02:00
1 parent 1866e38c25
commit cdc1cdc0cc
5 files changed
+63 -11

No files matched your search

+8 -4
View File
@@ -166,15 +166,19 @@ export class Prober {
const want = coordinatorConfig.probe.batch - this.queue.length
if (want <= 0) return
// The starting points chosen here: in the query Postgres drew random() once for all slices.
const [{ lo, hi }] = (await db.rawQuery(`SELECT min(id) AS lo, max(id) AS hi FROM probe_candidates`)).rows
// Over the claimable ones of this tier only: the rest of the queue isn't settled yet.
const [{ lo, hi }] = (
await db.rawQuery(`SELECT min(id) AS lo, max(id) AS hi FROM probe_candidates WHERE status = 'waiting' AND ready AND kind = ? AND unreadable >= ?`, [kind, unreadable])
).rows
if (lo === null) continue
const starts = Array.from({ length: SLICES }, () => Number(lo) + Math.floor(Math.random() * (Number(hi) - Number(lo) + 1)))
const { rows } = await db.rawQuery(
`UPDATE probe_candidates SET status = 'taken' WHERE id IN (
SELECT pc.id FROM unnest(?::bigint[]) AS o(start)
CROSS JOIN LATERAL (
SELECT pc.id FROM probe_candidates pc JOIN releases r ON r.id = pc.release_id
WHERE pc.status = 'waiting' AND pc.kind = ? AND r.settled_at IS NOT NULL AND (NOT ? OR NOT r.readable) AND pc.id >= o.start
ORDER BY pc.id LIMIT ? FOR UPDATE OF pc SKIP LOCKED) pc)
SELECT pc.id FROM probe_candidates pc
WHERE pc.status = 'waiting' AND pc.ready AND pc.kind = ? AND pc.unreadable >= ? AND pc.id >= o.start
ORDER BY pc.kind, pc.unreadable, pc.id LIMIT ? FOR UPDATE SKIP LOCKED) pc)
RETURNING id, release_id, kind, message_id`,
[starts, kind, unreadable, Math.ceil(want / SLICES)]
)
+12 -6
View File
@@ -21,14 +21,20 @@ class Settler {
async run(): Promise<number> {
let settled = 0
while (true) {
const { rowCount } = await db.rawQuery(
`UPDATE releases SET settled_at = now(), complete = CASE WHEN file_total IS NULL THEN NULL ELSE files >= file_total END
WHERE id IN (SELECT id FROM releases WHERE settled_at IS NULL AND last_seen_at < now() - make_interval(mins => ?)
AND (last_seen_at < now() - make_interval(hours => ?) OR coalesce(last_posted_at, posted_at) < now() - make_interval(hours => ?)) LIMIT ?)`,
// Their probe candidates become claimable (Prober.claim), the unreadable ones first.
const { rows } = await db.rawQuery(
`WITH settled AS (
UPDATE releases SET settled_at = now(), complete = CASE WHEN file_total IS NULL THEN NULL ELSE files >= file_total END
WHERE id IN (SELECT id FROM releases WHERE settled_at IS NULL AND last_seen_at < now() - make_interval(mins => ?)
AND (last_seen_at < now() - make_interval(hours => ?) OR coalesce(last_posted_at, posted_at) < now() - make_interval(hours => ?)) LIMIT ?)
RETURNING id, readable),
marked AS (UPDATE probe_candidates pc SET ready = true, unreadable = NOT s.readable FROM settled s WHERE pc.release_id = s.id)
SELECT count(*)::int AS n FROM settled`,
[QUIET_HISTORY_MINUTES, QUIET_HOURS, QUIET_HOURS, BATCH]
)
settled += rowCount ?? 0
if ((rowCount ?? 0) < BATCH) return settled
const n = rows[0].n
settled += n
if (n < BATCH) return settled
}
}
}
@@ -0,0 +1,30 @@
import { BaseSchema } from '@adonisjs/lucid/schema'
/**
* Candidates the prober may claim (Prober.claim): `ready` once their release
* settled (the settler sets it), `unreadable` when that release had no
* readable name then. The claims joined releases for both and started at
* random points of the queue; in its newer half nothing had settled, so a
* slice scanned to the end (0.3 s a slice, 20 slices, 5 tiers: the prober
* stood still for minutes per claim; LOG.md 2026-10-12 (20)). With an index
* over the claimable rows only.
*/
export default class extends BaseSchema {
static disableTransactions = true
async up() {
this.schema.raw(`ALTER TABLE probe_candidates ADD COLUMN IF NOT EXISTS ready boolean NOT NULL DEFAULT false`)
this.schema.raw(`ALTER TABLE probe_candidates ADD COLUMN IF NOT EXISTS unreadable boolean NOT NULL DEFAULT false`)
this.schema.raw(
`UPDATE probe_candidates pc SET ready = true, unreadable = NOT r.readable FROM releases r
WHERE r.id = pc.release_id AND r.settled_at IS NOT NULL AND NOT pc.ready`
)
this.schema.raw(`CREATE INDEX CONCURRENTLY IF NOT EXISTS probe_candidates_ready_idx ON probe_candidates (kind, unreadable, id) WHERE status = 'waiting' AND ready`)
this.schema.raw(`DROP INDEX CONCURRENTLY IF EXISTS probe_candidates_kind_waiting_idx`)
}
async down() {
this.schema.raw(`DROP INDEX CONCURRENTLY IF EXISTS probe_candidates_ready_idx`)
this.schema.raw(`ALTER TABLE probe_candidates DROP COLUMN IF EXISTS ready, DROP COLUMN IF EXISTS unreadable`)
}
}
+5 -1
View File
@@ -169,7 +169,7 @@ export class ProbeCacheSchema extends BaseModel {
}
export class ProbeCandidateSchema extends BaseModel {
static $columns = ['createdAt', 'file', 'id', 'kind', 'messageId', 'releaseId', 'status'] as const
static $columns = ['createdAt', 'file', 'id', 'kind', 'messageId', 'ready', 'releaseId', 'status', 'unreadable'] as const
$columns = ProbeCandidateSchema.$columns
@column.dateTime({ autoCreate: true })
declare createdAt: DateTime
@@ -182,9 +182,13 @@ export class ProbeCandidateSchema extends BaseModel {
@column()
declare messageId: string
@column()
declare ready: boolean
@column()
declare releaseId: bigint | number
@column()
declare status: string
@column()
declare unreadable: boolean
}
export class RangeSchema extends BaseModel {
+8
View File
@@ -24,7 +24,15 @@ test.group('Settler', (group) => {
})
const minutesAgo = (m: number) => new Date(Date.now() - m * 60_000)
await db.table('releases').multiInsert([row('old', minutesAgo(31), posted), row('fresh', minutesAgo(5), posted), row('tail', minutesAgo(31), minutesAgo(40))])
const ids = Object.fromEntries((await db.from('releases').select('id', 'name')).map((r) => [r.name, r.id]))
await db.table('probe_candidates').multiInsert([
{ release_id: ids.old, kind: 'lone', message_id: '<a@x>', file: 'a' },
{ release_id: ids.fresh, kind: 'lone', message_id: '<b@x>', file: 'b' },
])
assert.equal(await Settler.run(), 1)
// The settled release's candidates are claimable now, as unreadable (readable defaults to false).
const ready = await db.from('probe_candidates').orderBy('message_id').select('ready', 'unreadable')
assert.deepEqual(ready.map((r) => [r.ready, r.unreadable]), [[true, true], [false, false]])
const settled = (await db.from('releases').whereNotNull('settled_at').select('name')).map((r) => r.name)
assert.deepEqual(settled, ['old'])
})