Screenshots — sluice

Where are the credentials? The command lines below don't show any — by design. sluice reads connection DSNs from the SLUICE_SOURCE / SLUICE_TARGET environment variables (or a config file passed with --config), and secrets such as an encryption passphrase from --encryption-passphrase-env / --encryption-passphrase-file rather than a literal on the command line. Every command shown is exactly what runs — only the credentials are supplied out-of-band, so they never land in shell history or a screen recording.

`sync start`

A continuous-sync stream as a live status panel — mode, position, freshness, rows applied, throughput, health, and recent events.

CDC engagedSnapshot done; streaming live changes, position advancing.

ApplyingRows applied and throughput climb as source changes arrive.

Drain & stopq / ctrl+c drains in-flight changes, then exits cleanly.

`migrate`

A one-shot schema-and-data copy with a live phase checklist — tables, a per-table bulk-copy bar, then deferred indexes and constraints.

PhasesTables created; bulk copy begins, indexes and constraints deferred.

CompleteSummary panel: tables, rows, duration, and any warnings.

`verify`

Reconcile the target against the source, row-for-row (or by count), and report exactly what matched.

CheckingSchema compared, then the data comparison runs.

CleanTables checked / clean / mismatched / skipped.

`backup full`

A chunked, optionally encrypted and signed logical backup — with an EndPosition so a later incremental resumes exactly where it left off.

CopyingSchema captured; rows stream out in size-bounded chunks.

`restore`

Replay a backup chain into a fresh target — schema, data, then constraints.

`sync tui`

The fleet dashboard in your terminal — every stream's state, restarts, and time-in-state, with a detail pane for the selected stream.

`sync run --dashboard-listen`

The same fleet status as a browser dashboard, served on a local port — no terminal required.

Web viewThe embedded HTML dashboard at --dashboard-listen.

`sync health`

A scriptable per-stream health probe — freshness, position, lag, and spill — exit-code and JSON friendly for cron and CI.

`metrics-watch`

Live PlanetScale target metrics as a status panel — CPU, memory, storage, lag, and connections, sampled on a cadence, with threshold breaches surfaced as live events.

WatchingThe live metrics panel; breaches appear in the events region.

`notifications`

Threshold-breach alerts as a rendered HTML email (or Slack / webhook) — advisory, failure-isolated, never affecting the sync.

Email alertThe HTML email template for a storage-utilisation breach.

`cutover`

Prime the target's sequences and AUTO_INCREMENT columns past the migrated data so the app can write safely after the switch. Runs in milliseconds — here's the result panel.