Incident history
Every incident, written in plain words: what happened, who was affected, and what we changed so it does not repeat.
Degraded
12 July 2026 · 04:10–04:52 UTC · resolved
Runsheet loading was slow for some vendors.
Runsheets took up to 20 seconds to open for about 8% of vendors during the morning window. Deliveries were not blocked, and printed backups were unaffected. Cause: a slow database query after a routine update. Fix: the query was corrected and we added a check that catches this class of slowdown before release.
Degraded
28 May 2026 · 21:34–21:58 UTC · resolved
Wallet top-ups were delayed by a payment partner.
Top-ups made through one payment method took up to 15 minutes to appear in wallets. No money was lost or double-counted; the ledger reconciled every delayed top-up automatically once the partner recovered. Fix: we now show customers a clear "top-up on the way" state instead of silence.
Maintenance
3 May 2026 · 22:00–22:20 UTC · completed
Planned database maintenance.
Scheduled ahead of time and placed outside every active region's delivery window. The nightly forecast ran normally after the window closed. No action was needed from vendors.
[placeholder incidents: replace with the real history before launch]