Continue a failed import from the step that failed (Migration.resume), with the project, databases and variables it already created: nothing finished runs again. Fix the cause first; a failed preview check is deployed and checked again (acceptPreview: continue as it is)

the automatic check of the preview: broken links and files, mixed content, old site vs preview, WordPress login and old addresses

cron jobs of the old host created on the project when the import finished (paused ones: POST /migrations/{id}/crons/enable, or the project's cron jobs)

vercel: what happened to the external databases, Vercel Blob and vercel.json routing

vercel/heroku with moved databases: their final copy right before the domain switches

cpanel: the account's mail moved to Hostingsurge's mail service (POST /migrations/{id}/mail)

preview_ready, cutover and live: whether the old hosting can be cancelled, and what still lives there

exactly which values of envLater were left out, per environment (the name's other values moved with the import)

after the start: every variable of the old host and where it went, or exactly why not

vercel: databases whose connection string was read with POST /migrations/{id}/databases/{key}/connect (never the password)

never the values themselves

failed: the import can continue from the failed step (POST /migrations/{id}/retry) once the cause is fixed

failed or canceled: what POST /migrations/{id}/discard deletes before you start over

the import is queued, not running: one import runs at a time per organization, and this one starts by itself once the one running now is done or waits for you. Null while it runs or waits for you