Monitoring Azure Backup workloads
Azure Backup provides multiple backup solutions based on the backup requirement and infrastructure topology (On-premises vs Azure). Any backup user or admin should see what is going on across all solutions and expected to be notified in important scenarios. This article details the monitoring and notification capabilities provided by Azure Backup service.
Backup Jobs in Recovery Services vault
Azure Backup provides in-built monitoring and alerting capabilities for workloads being protected by Azure Backup. In the Recovery Services vault settings, the Monitoring section provides in-built jobs and alerts.
Jobs are generated when operations such as configuring backup, backup, restore, delete backup, and so on, are performed.
Jobs from the following Azure Backup solutions are shown here:
- Azure VM backup
- Azure File backup
- Azure workload backup such as SQL
- Azure Backup agent (MAB)
Jobs from System Center Data Protection Manager (SC-DPM), Microsoft Azure Backup Server (MABS) are NOT displayed.
Azure workloads such as SQL backups within Azure VMs have huge number of backup jobs. For example, log backups can run for every 15 minutes. Hence, for such DB workloads, only user triggered operations are displayed. Scheduled backup operations are NOT displayed.
Backup Alerts in Recovery Services vault
Alerts are primarily scenarios where users are notified so that they can take relevant action. The Backup Alerts section shows alerts generated by Azure Backup service. These alerts are defined by the service and user cannot custom create any alerts.
The following scenarios are defined by service as alertable scenarios.
- Backup/Restore failures
- Backup succeeded with warnings for Azure Backup Agent (MAB)
- Stop protection with retain data/Stop protection with delete data
Exceptions when an alert is not raised
There are few exceptions when an alert is not raised on a failure, they are:
- User explicitly canceled the running job
- The job fails because another backup job is in progress (nothing to act on here since we just have to wait for the previous job to finish)
- The VM backup job fails because the backed-up Azure VM no longer exists
The above exceptions are designed from the understanding that the result of these operations (primarily user triggered) shows up immediately on portal/PS/CLI clients. Hence, the user is immediately aware and doesn't need a notification.
Alerts from the following Azure Backup solutions are shown here:
- Azure VM backups
- Azure File backups
- Azure workload backups such as SQL
- Azure Backup agent (MAB)
Alerts from System Center Data Protection Manager (SC-DPM), Microsoft Azure Backup Server (MABS) are NOT displayed here.
Based on alert severity, alerts can be defined in three types:
- Critical: In principle, any backup or recovery failure (scheduled or user triggered) would lead to generation of an alert and would be shown as a Critical alert and also destructive operations such as delete backup.
- Warning: If the backup operation succeeds but with few warnings, they are listed as Warning alerts.
- Informational: As of today, no informational alert is generated by Azure Backup service.
Notification for Backup Alerts
Configuration of notification can be done only through Azure Portal. PS/CLI/REST API/Azure Resource Manager Template support is not supported.
Once an alert is raised, users are notified. Azure Backup provides an inbuilt notification mechanism via e-mail. One can specify individual email addresses or distribution lists to be notified when an alert is generated. You can also choose whether to get notified for each individual alert or to group them in an hourly digest and then get notified.
The alerts for SQL backups will be consolidated and the email is sent only for the first occurrence. But if the alert is inactivated by the user, the next occurrence will trigger another email.
When notification is configured, you will receive a welcome or introductory email. This confirms that Azure Backup can send emails to these addresses when an alert is raised.
If the frequency was set to an hourly digest and an alert was raised and resolved within an hour, it will not be a part of the upcoming hourly digest.
- If a destructive operation such as stop protection with delete data is performed, an alert is raised and an email is sent to subscription owners, admins, and co-admins even if notifications are NOT configured for the Recover Service vault.
- To configure notification for successful jobs use Log Analytics.
To inactivate/resolve an active alert, you can click on the list item corresponding to the alert you wish to inactivate. This opens up a screen which displays detailed information about the alert, with an 'Inactivate' button on the top. Clicking this button would change the status of the alert to 'Inactive'. You may also inactivate an alert by simply right-clicking on the list item corresponding to that alert and selecting 'Inactivate'.