FailedFailed 2 times in a row
Backup job "DC-02 · vzdump all VMs" reported failed: [Failed] vzdump all VMs — repository unreachable
One list of what needs someone, worst first.
Needs attention collects every backup that failed, went missing or finished with warnings, plus the report emails nobody could read and the alert channels that stopped delivering. Work down it in the morning and the estate is checked.
Everything that can stop a backup being trusted.
Each item says what happened in plain words. Failed, missing and warning backups are alerts: one per backup, with a state. The others clear by themselves once the cause is gone.
- Failed
- “Backup failed”, or “Failed 4 times in a row” when the failures repeat. A Synology integrity check that fails shows as “Integrity check failed”.
- Missing
- “Report missing”: the report did not arrive inside its window.
- Warning
- “Completed with warnings”, or “Integrity check warning”.
- Size drop
- “Backup shrank 62%”: a backup that reported success but wrote half its usual size or less. The Expected button hides a drop you planned.
- Report emails
- “3 emails with no readable result” or “2 emails not matched to a backup”, from the last 7 days, grouped per client.
- Alert channels
- “Alert channel failing” when a delivery failed in the last 24 hours, and “No alert channel set up” once the first backup exists.
The worst problem is always at the top.
The list is sorted by priority, so the first row is the one to fix first. Open items come before the ones someone has already picked up, and a failed backup before a warning.
Sort by Oldest first or Newest first instead, group the list by client, filter to one client, or search it.
- State: open, then acknowledged, then snoozed, then resolved.
- Severity: critical before warning.
- Kind: failed, missing, a failing alert channel, no alert channel, warning, size drop, report emails.
- Age: newest first.
Everyone can see who has it, and what fixed it last time.
Resolving an item takes a note of what was done, up to 500 characters. The next time the same backup has a problem, its row shows the latest one as Fixed before:, and the resolve form suggests up to four earlier notes under Fixed before on this backup.
Owners, admins and members can work the list. Viewers can read it, move through it and open items.
- Acknowledge
- Records who and when. The item stays listed until the backup recovers.
- Snooze
- Until the next report is due, 1 hour, 4 hours, until tomorrow 08:00 or 24 hours. When it ends the item is open again.
- Resolve
- With a note. The channels that received the alert are told who resolved it.
- Assign
- To a teammate. Filter the list by Anyone, Assigned to me or Unassigned.
Clear a long morning list from the keyboard.
Select several items and acknowledge or resolve them together; one note covers all of them. Acknowledge all takes every open problem, and Acknowledge all for this client takes one client’s items when the list is grouped.
- Bulk actions: up to 200 items at a time.
- Export: the list as CSV, for a ticket or a handover.
| Key | What it does |
|---|---|
| j / k | Next / previous problem (the arrow keys work too) |
| a | Acknowledge |
| s | Snooze until the next report (or 1 hour) |
| r | Resolve with a note |
| x | Select or unselect |
| o or Enter | Open the backup |
| ? | Show or hide the shortcuts |
A backup that recovers takes itself off the list.
Nobody has to close an alert the backup has already fixed. The next good report does it, and says so where the alert was sent.
- An OK report arrives for a backup with an open alert.
- The alert resolves, and the item moves to the Resolved tab.
- A recovery notice, “Recovered: <job> is healthy again”, goes to the channels that delivered the alert.
- Resolved by hand instead, the same channels get “Resolved by <name>”.
What it does not do.
- Only alerts (failed, missing and warning backups) can be acknowledged, snoozed or resolved. Report emails, channel problems and size drops clear when their cause does.
- A snooze lasts between 15 minutes and 7 days. When it ends the item is open again, but nothing is sent a second time.
- Size drops are listed here only. They are never sent to alert channels.
- A bulk action covers at most 200 items, and the Resolved tab shows the latest 100.
Start the morning with one list.
Point one client's report emails at BackupSentinel and its failed, missing and warning backups show up here.