Availability: Migrate isn't offered for purchase or activation at launch. It's coming later. This page explains how Migrate works so you can plan client moves ahead of time. There's no Migrate pricing yet. To hear when it's available, submit a request.
Once a job is configured and an agent is online, you can start the migration. Migrate tracks progress in real time and provides controls to pause, cancel, or re-run a job. This article covers starting a job, watching it run, and handling errors.
Before You Start
Confirm:
- The agent assigned to this job is showing as Online in the Agents list.
- Credentials for both source and destination are in the vault and attached to the job.
- You have reviewed the job's filter and configuration settings.
- If this is a destructive job (delete after upload), narrow it first: no scan can be run to preview what will be touched (see the correction note at the top of this page), so scope the job with its source filter, review its filter and configuration, and test a small set before you enable the delete.
Starting a Job
- Open the job and click Start (or use the Jobs list action menu).
- The job moves to Queued status. The agent picks it up on its next command poll (commands expire after 24 hours if not picked up).
- Status changes to Running once the agent begins transferring data.
Monitoring Progress
- Open the job and click Progress.
- The progress view shows real-time counters: files scanned, files uploaded, files verified, files errored, and bytes transferred.
- Progress is ingested in batches. Each batch has a unique ID — if the agent retries a batch after a network interruption, the duplicate is detected and the counters are not inflated.
Stall Detection
If a running job records no progress for 20 consecutive minutes (and has been running for at least 5 minutes), Migrate raises a stall event and writes it to the job's audit trail. Migrate does not message you about the stall. The same job will not raise another stall event for at least 2 hours, so you will not be flooded with repeats for a single slow job.
If you see a stall, check:
- The agent is still Online (heartbeats may have stopped).
- The source system is reachable.
- Bandwidth limits are not set too low.
Job Controls
| Control | What It Does |
|---|---|
| Pause | Stops the transfer; the job is left Partial. There is no Resume — use Rerun to continue from the checkpoint |
| Cancel | Stops the job permanently; staging data is retained for 7 days then cleaned up |
| Rerun | Queues the job again; it resumes from the job's saved checkpoint — items already transferred are skipped and items that failed are retried |
| Rollback | Not implemented — the control is disabled in the console and the API refuses rollback_job with 501 not_implemented. Migrate cannot automatically reverse a migration; see Rollback
|
Error Handling
Files that fail to upload are recorded as errors in the progress batch. The job continues processing remaining files — a single file error does not stop the job. After the job completes or fails, check the progress detail and error log to see which items need attention.
If the job fails entirely, check:
- Agent went offline mid-job. The offline checker flips the agent status after the heartbeat threshold. When the agent comes back online, use Rerun to restart.
- Credential fetch failed. The agent may have hit the 10 fetches/hour rate limit, or the credential IDs do not match the job. Check the audit log.
- Staging URL errors. R2 credentials must be configured. Contact your admin if staging URLs are failing.
FAQ
Can I run multiple jobs at the same time? An agent can only handle one job at a time (status: Busy). If you have multiple jobs to run in parallel, install agents on separate machines.
What happens if the agent disconnects mid-migration? The heartbeat stops and the agent is marked Offline after the threshold period. The job pauses. When the agent reconnects, use Rerun — it resumes from the job's saved checkpoint, so items already transferred are skipped and items that failed are retried.
How do I know a job is truly complete? Status moves to Completed. The progress counters show zero errored files if everything transferred cleanly. Generate a migration report for the formal record.