Interruptions and recovery
Outcome: you will know what every interruption means, what is at risk in each one (usually nothing), and how to resume.
The first thing to know
Section titled “The first thing to know”Nothing recorded on the chain is affected by any outage of this product. Your artifacts, balances, and transactions live in Zcash blocks; the indexer only reads them. An outage delays what the product can show you and changes nothing about what you own.
When the site cannot reach the indexer
Section titled “When the site cannot reach the indexer”The page tells you once, keeps what it can still show, and fixes itself when the service comes back:
- One notice at the top says what is unavailable and what is not; each affected section is marked with one line rather than repeating the explanation.
- Anything already on screen stays on screen, labeled as last known.
- The page retries on its own, less often the longer an outage lasts so it does not add load to a struggling service, and fills itself back in when the indexer answers. You do not have to reload. The notice carries a Retry now button if you would rather not wait.
- While the site cannot confirm current chain data, creating, etching, minting, and transferring stay unavailable rather than proceeding on stale information.
When data is stale
Section titled “When data is stale”Status information (node height, indexer height, lag, mempool state) is always shown honestly. Stale data is labeled STALE with the time it was last confirmed, and last-known values stay visible rather than vanishing.
When the chain reorganizes
Section titled “When the chain reorganizes”The indexer journals every event with its block height and hash. When the chain reorganizes, everything recomputes deterministically from the surviving chain: inscriptions in orphaned blocks revert to pending or disappear exactly as chain state dictates, and sequence numbers and ownership are recalculated. A reorganization deeper than the supported automatic bound stops the indexer for operator repair rather than serving wrong data.
This is why sequence numbers are stable only after confirmation depth, and why fresh inscriptions display their provisional nature.
Resuming interrupted orders
Section titled “Resuming interrupted orders”Inscribe orders and batches persist in your browser on this device:
- Declined commit: nothing was signed or broadcast; the order is cancelled cleanly.
- Declined or interrupted reveal: the commit is already on chain; the order stays open and resumable from its timeline.
- Closed tab: reopen Inscribe; prepared orders and batches are still there.
Each order’s timeline shows real chain states, so on resume you see exactly where it stopped and what the next approval is.
How to verify after any interruption
Section titled “How to verify after any interruption”- Check the status page for the last recorded coverage and freshness, or the order’s own page for where it stands now.
- Search the relevant txid or address in Scan; the chain record is the authority, and it was never at risk.
Related
Section titled “Related”Recovery list: candidate update
Section titled “Recovery list: candidate update”This update awaits release qualification. Saved orders and creation-page resume lists show readable states and the next action. In the candidate, creation, draft editing and recovery stay accessible during read failures. Submission depends on the checks required by that operation. Unknown availability or funding safety is never treated as permission to spend. Open the original order to review, sign, pay or inspect a payment difference. Opening a list never pays or retries an order.
A delayed update keeps the last verified state with a brief stale marker. The original recovery link remains the route back to that order. Do not create or pay a replacement invoice because a status read timed out.
Portfolio sections load when selected. A partial listing lookup keeps unread listings unknown; it does not mark them unlisted.