Create, verify, and delete a time-based autoscaling schedule for a ClickHouse Cloud service using the scalingSchedule beta API

Documented steps

1. Check for an existing schedule: GET /v1/organizations/{organizationId}/services/{serviceId}/scalingSchedule. It returns {entries:[...], baseConfig:{...}, activeEntryId} when configured, or 404 when none exists.
2. Each entry requires name, weekdays, startHourUtc, and endHourUtc. weekdays is an array where 0 = Sunday through 6 = Saturday; startHourUtc is 0-23 inclusive; endHourUtc is 1-24 exclusive (use 24 for midnight). There is no cron syntax.
3. Optional per-entry fields: autoscalingMode ('vertical' default, or 'horizontal'), minReplicaMemoryGb/maxReplicaMemoryGb (multiples of 4, 8-356), numReplicas, minReplicas/maxReplicas, idleScaling, idleTimeoutMinutes.
4. Create or fully replace the schedule with POST /v1/organizations/{organizationId}/services/{serviceId}/scalingSchedule and body {"entries":[...]}. POST always overwrites the entire entry list - it is a replace, not a merge.
5. curl -u "$KEY_ID:$KEY_SECRET" -X POST https://api.clickhouse.cloud/v1/organizations/$ORG_ID/services/$SERVICE_ID/scalingSchedule -H 'Content-Type: application/json' -d '{"entries":[{"name":"Business hours","weekdays":[1,2,3,4,5],"startHourUtc":9,"endHourUtc":17,"autoscalingMode":"vertical","minReplicaMemoryGb":16,"maxReplicaMemoryGb":64}]}'
6. For a horizontal entry set autoscalingMode:"horizontal", make minReplicaMemoryGb equal maxReplicaMemoryGb, and supply a minReplicas/maxReplicas band instead of numReplicas (the two are mutually exclusive).
7. The base configuration applied outside any window is NOT set here - it is the service's normal scaling config managed by PATCH .../replicaScaling. scalingSchedule only layers time-boxed overrides on top of it.
8. Verify by re-reading the schedule and checking activeEntryId (the currently active entry, absent when the base config is in effect), then GET the service itself to confirm live memory/replica values, since scheduled changes apply best-effort and can lag a few minutes.
9. Clear the schedule with DELETE /v1/organizations/{organizationId}/services/{serviceId}/scalingSchedule, or by POSTing {"entries":[]}. A subsequent GET then returns 404.
10. Reference: https://clickhouse.com/docs/cloud/manage/api/swagger and https://clickhouse.com/docs/cloud/manage/api/api-overview