fix(scheduler): decrease tasks table autovacuum factor settings

The performance of the expireTasks query decreases over time until
vacuuming occurs and speed of the query is back to its baseline.
The tasks table update rate is very high (ex: frequent state and dates updates)
and create table bloat.
This commit is an attempts at optimizing the autovacuum settings for the
tasks table to trigger cleaning operations more frequently by reducing
scale factors. (The pg default autovaccum settings are too high for the tasks table)
By vacuuming at 1% modified rows rather than the default 10%, we'll maintain more
consistent query performance while preventing excessive dead tuple accumulation
I choose the new value somehow arbitrarely. I will keep monitoring the scheduler queries

Author: TBonnin

packages/scheduler/lib/db/migrations/20250509143200_tasks_autovacuum_settings.ts

@@ -0,0 +1,14 @@
+import type { Knex } from 'knex';
+import { TASKS_TABLE } from '../../models/tasks.js';
+
+export async function up(knex: Knex): Promise<void> {
+    await knex.raw(`
+            ALTER TABLE ${TASKS_TABLE} SET (
+                autovacuum_vacuum_scale_factor = 0.01,        -- reducing from 10% to 1%
+                autovacuum_analyze_scale_factor = 0.01,       -- reducing from 5% to 1%
+                autovacuum_vacuum_insert_scale_factor = 0.05  -- reducing from 20% to 5%
+            );
+        `);
+}
+
+export async function down(): Promise<void> {}