Job-wide Auto Tune backoff/notification config

Job-wide backoff policy — the "when a change misbehaves" guardrail for every enabled optimization on the job.

Stored under the backoff_policy key in databricks_auto_tune_configuration.configuration, the same slot the cluster port uses, but only with the one setting a job has: how far a run's latency may drift from its pre-tune baseline.

The cluster policy's churn / cold-start counters, its cold_start_threshold_pct and its trigger_action are deliberately absent. Each describes behaviour the cluster and warehouse backoff monitors implement and nothing on the job side evaluates, and the jobs drawer renders none of them. Keys left in a stored row from before this split are ignored on read and dropped on the next save.

Cluster-wide Auto Tune notification preferences.

Stored under the notification_config key in databricks_auto_tune_configuration.configuration. References Slack channel ids from the tenant's notifications table by reference only — never a webhook URL inline (see NotificationConfigPublic for the same redaction rule on the generic notifications settings page).