The backup system of a working hosting provider: seven tiers across three independent locations, every number reported by the jobs themselves. Names, locations and providers are anonymized.
The Saturday pull from offsite provider A was cut off part-way. The partial copy was not kept as a restore point. An automatic retry on Sunday completed in full and was verified.
The scheduled pulls on 12 and 13 October did not start. A manual run on the 13th failed final verification on one file and was rejected. The next scheduled run on the 14th succeeded. All other tiers ran normally throughout.
The weekly upload was interrupted on the provider’s side. The next night’s upload completed automatically. The same copy existed at provider A the whole time.
Failed copies are never kept as restore points, and they are never hidden from this page.
No restore test has been recorded yet, so the rule's last digit, zero errors on restore, isn't proven. Two tiers are still filling up to their retention target. Both show on the page until they're fixed; neither is hidden behind a green light.
No number is typed by hand. Each job reports to Servatum when it finishes and the page is rebuilt every 15 minutes. If the data is more than two hours old, the page says so at the top.