Status
All systems operational · 2 issues ongoing
Checked live when you loaded this page: Sun, 18 Oct 2026 02:18:47 GMT. Reload to check again. Machine-readable: /api/health.
- Operational
App
This page rendered, so the app is serving.
- Operational
Database
Answered in 781 ms.
- Operational
Schema
Every migration this build shipped with is applied.
- Operational
Customer path
A test customer signed in, walked an inspection, added a photo, completed it, opened the share link and downloaded the PDF on a phone — last walked Sat, 17 Oct 2026 20:56:36 GMT. It is meant to run every two hours on GitHub Actions; while that account is stopped it runs when we start it by hand, so the date above is the truth, not the schedule.
- Operational
Mail
Deliveries confirmed by the provider in the last day.
- Operational
Payments
Stripe's webhook endpoint for this site is switched on, and every event Stripe has sent has arrived. Checked Sun, 18 Oct 2026 02:10:26 GMT.
- Operational
Backups
The last completed copy to the second provider finished Sat, 17 Oct 2026 17:08:10 GMT.
Incident log
Kept since 2026-09-01. Each incident is dated from when customers were first affected, which can be before it was found.
2026-09-26
Nightly backups stopped (ongoing)
The nightly copy of the database and stored files to the second provider did not run from 25 September to 4 October: the scheduled job runs on GitHub Actions, which stopped running every workflow on the account. A full copy was taken by hand on 4 October and read back from the second provider. Live data is unaffected; what is at risk is what changes between copies, if the primary host were lost. The ten-minute check reads the backup receipt and says when it is behind, and /status shows the date of the last completed copy. Open until copies land nightly again on their own.
2026-09-24
Email and the sitesui.app addresses went dark
sitesui.app, the domain that carried hello@, support@ and the app's sending address, was put on hold at the registrar on 24 September. From the next morning outbound mail — receipts, verification links, share links, vendor jobs — was refused. Sending moved to mail.sitesui.com on 26 September and mail has gone out since; the hold on sitesui.app lifted on 27 September.
2026-09-22 · 4320 min
Nightly backups missed three nights
The copy to the second provider did not run on 22, 23 or 24 September and nothing noticed; it was found on the 25th when the bucket was listed by hand. The ten-minute check now reads the last backup receipt and raises an alert when a night's copy has not landed.
2026-09-21
AI suggestions on resident reports unavailable
The model that suggests a trade and an urgency for a resident's report stopped answering when the account's key could no longer be used, and the same key drafts our blog. Reports still arrive, are saved and reach the inbox; reports that mention gas, smoke, water or exposed wiring are flagged as safety risks and rise to the top without the model. The ten-minute check now reads whether the model can be called. Resolved 9 October: the product moved back to Anthropic's Claude on a working key, and the ten-minute check read every system healthy that evening.
2026-08-20
Work orders page failed for paid workspaces
The work-orders page returned an error for every paid workspace after a schema change reached the code before it reached the database. Found by hand on 1 September and fixed the same day. Since then the app applies its own migrations at boot and a health check outside the app watches for the gap, so a deploy cannot serve traffic against a database that lacks what it needs.
2026-03-06
Vendor texts not delivered (ongoing)
Work orders sent to a vendor's mobile number were refused by the carrier because our toll-free number's verification was rejected, while the app recorded them as queued. Found on 21 September. The app now says texts are paused and sends the job by email with a link the vendor answers from; the carrier's answer is read back every hour. Open until the refiled verification is approved.
Recovery, plainly
- Backups. The backup is designed to copy the full database and every stored file to a second provider each night, separate from the primary host. The last completed copy finished Sat, 17 Oct 2026 17:08:10 GMT. A rehearsal that restores a copy into a fresh database is run on demand; this page does not yet show when it last passed.
- Recovery point. The most that can be lost in a total failure of the primary host is what changed since the last completed copy, dated above. We do not yet publish a recovery-time target, because we have not rehearsed a full restore under the clock enough times to state one honestly; when we have, it goes here.
- Offline. A walk in progress keeps ratings, notes and photos on the phone and syncs when the app is back; an outage costs waiting, not the inspection.
- No uptime number yet. We will publish one when an outside monitor has a year of data to stand behind it.
- Report a problem. hello@sitesui.app; a person answers. Security disclosures: see Security.