1 — Select
Browse the configured production bucket and select exact keys or prefixes.
s3://lab-prod/runs/2026-06/One controlled path from hot data to cold storage—and back.
Data Convoy turns S3 archiving into a visible, permissioned workflow. Teams select real objects, preview the work, send it to customer-operated workers, and keep a durable record of the result.
The workflow
Browse. Preview. Archive. Verify.
Browse a team’s production bucket, choose objects or prefixes, and run a dry preview before creating work. Archive jobs copy into Glacier or Glacier Deep Archive. When verify-and-remove is enabled, the worker checks integrity before deleting the production source.
2 — Preview
Dry runs crawl the selection, apply exclusion rules, and report counts and bytes without copying or rehydrating.
3 — Execute
A worker assigned to the team claims the queued job and operates in your AWS account.
4 — Prove
Checksums are compared when available, with size verification as a fallback. Results and failures remain attached to the job.
Restore is a first-class workflow
Restores issue Glacier rehydration requests in parallel, wait for objects to become available, and copy them back with the same checksum-then-size verification used for archive. A restore dry run shows scope before retrieval costs begin.
Version-aware
Restore handling is designed around real S3 object state rather than a local file-list abstraction.
Failure visibility
Per-object failures remain non-fatal, while any failed item makes the overall job fail so partial success is never mistaken for completion.
Cancel and retry
Operators can cancel queued or running work and retry eligible failures within a bounded retry policy.
Governance follows the data
Each team has its own buckets, members, workers, defaults, schedules, jobs, and audit view. Workers can serve several assigned teams, but claim authorization is re-checked against the owner’s current role on every job claim.
Your buckets. Your workers. Your control.