fix(dts): [173335400] pause running sync job before isolation on delete

[Rajendraladkat1919]

What
Fix tencentcloud_dts_sync_job to pause a running job before attempting isolation on delete, and return nil instead of an error when the job is not found during Read.

Why
Delete bug: The Delete function calls IsolateSyncJobs directly regardless of the job's current state. Tencent rejects isolation on a running or syncing job with:

UnsupportedOperation.UnsupportedOperationError: Current job is temporarily unable to respond to isolation events
This makes it impossible to destroy a tencentcloud_dts_sync_job resource through Terraform while the job is actively syncing — which is the normal production state. Every terraform destroy or config removal requires manual console intervention to pause the job first.

Read bug: When a job is not found, the Read function calls d.SetId("") correctly but then returns fmt.Errorf(...) instead of nil. The error overrides the state ID clear, so Terraform fails every subsequent plan with resource syncJob sync-xxx does not exist rather than automatically removing the stale entry from state. This breaks pipelines whenever a job is deleted outside Terraform.

Type of change
Bug fix
Changes
resource_tc_dts_sync_job.go — Delete:
Added a pre-isolation status check. If the job is in a running/syncing state, PauseSyncJob is called first and the function waits for Paused status before proceeding with isolation. Jobs already in NotBilledByInternational or NotBilled trade status are treated as gone and return nil immediately.

resource_tc_dts_sync_job.go — Read:
Changed return fmt.Errorf(...) to return nil after d.SetId("") so Terraform gracefully removes missing resources from state instead of erroring.

service_tencentcloud_dts.go:
Added PauseDtsSyncJobById helper following the same pattern as the existing IsolateDtsSyncJobById.

Checklist
make build succeeds locally
make fmt produces no diff
make lint passes
make check-mux passes
Acceptance tests added or updated
Changelog entry added under .changelog/.txt

Notes for reviewers
The delete sequence is now: check status → pause if running → wait for Paused → isolate → wait for Isolated/NotBilled → destroy → wait for Offlined. This mirrors the manual steps required to delete a DTS sync job via the Tencent Cloud Console and follows the same polling pattern already used elsewhere in this file.

The NotBilledByInternational early-exit path (lines 329–331 in the original) is preserved — PostPay jobs that enter this state after isolation are auto-removed by Tencent and do not need an explicit destroy call.

[Rajendraladkat1919]

Im an external contributor and don't have access to an internal Tencent ticket ID required by the PR title format ([{9-digit-id}]). Could a maintainer assign one and update the title before merging?