Page History

The Days since last backup alerts indicate the number of days that databases (per-database or at the instance level) have not been backed up.

To see more information about these alerts applied to Availability Groups, check Days since last backup for availability groups

Anchor
instancelevel instancelevel

Days since last backup at the instance-level

...

1. Right-click a SQL Server instance in the Servers tree.
2. Select Configure Alerts.
3. Select the Days since last backup metric in the Alert Configuration window.
4. If the metric features a per-database, on the Configuration tab of the Alert Configuration window, click Add. Use the drop-down list to select the database to which you want to apply these settings. 
5. Check the boxes next to Informational, Warning, and Critical to include alerts for these states.
6. Change the alert thresholds by moving the arrows to the appropriate levels or by double-clicking the value and typing a new threshold level.
7. Click Advanced to apply alert suppression.
8. Click OK to accept your changes.

Anchor
agmetric agmetric

Days since last backup for availability groups

Here are some escenarios to understand this metric relates to Availability Groups:

When primary and (any or all) other secondary servers are being monitored: Primary server will show the last backup date alert by checking all other secondary servers for the latest backup date of that database. If the latest backup date across that AG justifies the alert, it will come.

...

The secondary monitored databases will also show the same alert as the latest backup date for a AG db across all replicas will be the same

When only primary replica of a database is being monitored: The alert will come after checking the latest backup date across all replicas and if that date justifies an alert

When some secondary replica is being monitored but no primary replica: The alert will come after checking the latest backup date across all replicas and if that date justifies an alert

Limitations

1. We will have to assume that the connection credentials of all the replicas are the same. At this point SQLdm stores connection credentials of only the monitored servers. If a secondary replica is not monitored, we will not have its connection credentials. We cannot put this setting in the properties as number of secondary replicas is not limited and also its db specific.

2. A replica may have the latest backup date, but its possible that the sync is broken between replicas, in which case the db would have been backed up but the some data would not be there. In this case, the alert may not come if the last back update does not justify an alert.