TiDB provider adds scheduled backups with retention

TiDB backups no longer need someone to start them. The TiDB provider v0.1.5 adds cron schedules to the snapshot backups introduced in v0.1.3. Add schedules to a storage under spec.backup.storages[] and choose a retention: keep the last N successful backups, or everything from the last few days, weeks or months.

TiDB Operator v2 ships the BackupSchedule resource without a controller to run it, so the provider runs the schedules itself. Each cron slot creates a regular OpenEverest Backup that runs exactly like an on-demand one. A slot only fires while the instance is Ready or Updating and the previous run of the same schedule has finished. If it can’t start within 10 minutes, it’s skipped rather than queued.

Retention counts only successful backups, so a run of failures never pushes good backups out, and a time window always keeps the newest successful backup. An invalid cron expression shows up on the instance’s BackupConfigured condition. More details are in the provider README. Requires OpenEverest v2.0.0-dev.4 or later.