Extend upgrade_monitor_test.go to cover frozen phase job execution

**Epic:** #1080

Add test cases to `internal/daemon/consensus/upgrade_monitor_test.go`:
- Jobs dir present with one `(upgrade, frozen)` job → runs before upgrade scripts
- Jobs dir absent → upgrade scripts run normally (no-op)
- Abort failure → DaemonResult=Failed with `ReasonDaemonCustomJobFailed`
- Ignore failure → warning logged; upgrade scripts run
- Sequence order: 3 jobs at frozen phase run in ascending sequence

## Acceptance Criteria
- [ ] All scenarios covered with fake K8s client
- [ ] Tests are hermetic (no real cluster needed)
- [ ] Existing tests unaffected

Start with internal/daemon/consensus/upgrade_monitor_test.go and its existing fake Kubernetes client setup. Add coverage for frozen-phase job ordering, absent jobs, abort failures, and ignore failures, then run the focused test file. Done means every listed scenario is hermetic and the existing tests remain unaffected.

Written by the indexing model from the issue text.