System status and recovery record
Checked 23 Aug, 09:46 UK time, when you loaded this page. Nothing here is cached.
That line only tells you about this second. The part worth reading is further down: how often we copy your data, how long we keep it, and whether we have ever proved a copy can actually be restored.
Components
Backup and recovery record
Last 90 days: 82 days with backup runs, all of them successful.
A full database dump runs every hour. Each one is compressed, then encrypted with AES-256 before it leaves the server. Every night a separate job takes the newest backup and restores it into a throwaway database, which is the part most backup routines never do. It then checks that the restored copy has the same number of tables as the live system, and that workshops, users, jobs and invoices all came back with a sensible number of rows. If any of that disagrees, it emails us.
A note on the record before 14 August 2026. Because the hourly job writes one file per day, every run after the first was refused by the retention lock that protects finished backups from being overwritten. Those refusals were logged as failures, which they were not: the day’s backup was already stored and protected each time. They are still counted as runs here, no longer as failures, and the job stopped producing them on 14 August 2026. A backup that genuinely fails to reach the bucket still counts against us in full.
10 August 2026, on our nightly restore test. The job reported: validate: table count mismatch: restored 91 vs live 92. The next run succeeded 24 hours later. The 14 runs since have all passed.
Incidents
No incidents have been recorded. We publish one here for any unplanned outage, and a write-up of what happened for anything lasting more than an hour.
What you can check yourself
We hold no ISO 27001, no SOC 2 and no Cyber Essentials certification, and we would rather write that here than leave a gap where you assume one. Autera is built and run by a small UK team. What we offer instead is the record above and the documents beside it, all of which you can read without speaking to us first.
If something goes wrong
We post here first, and we email you if the problem affects your data rather than just your afternoon. If you think something is broken and this page says otherwise, tell us and we will go and look, because that gap is the most useful message we get. [email protected]
Every figure in the recovery record above is counted from our own backup log when you load this page, not written in by hand. Retention is our policy rather than a count, and the delivery figure comes from our email provider’s log.