NAKIVO Backup & Replication

The NAKIVO Backup & Replication check reads the PostgreSQL database of the NAKIVO Director read-only. Per job it reports the result of the last run, the age of the last successful run, long and skipped runs, failing objects (VMs), and the state and free space of the backup repositories. The check runs every 60 minutes on the octoja agent on Windows, Linux, and macOS (each as an x64 and an Arm64 build). It connects to PostgreSQL over the network, so it can run on any agent device that can reach the database.

Prerequisites

• The octoja agent can reach the NAKIVO Director's PostgreSQL database on its host and port. When the check runs on the Director itself, use 127.0.0.1.
• A dedicated database role with read access to the NAKIVO database. The check reads the tables jobs, jobexecutions, jobexecutionsummary, jobexecutionsteps, backuprepositories, backuprepositorystateinfo, backuprepositorystats and savepoints.
• The Monitoring Check Management permission to add the check.

Steps

Add the check on the device that can reach the database — usually the Director itself.

1. Open the device under Devices.
2. Select Checks → Add Check → NAKIVO Backup & Replication.

For several Directors, you can assign it with config packages.

Settings

The following fields appear only after you turn on Show advanced options:

If the warning and critical thresholds are entered the wrong way round, the check swaps them before evaluating.

Results

Each monitored job and each repository is evaluated on its own. The worst individual result becomes the check status.

The age thresholds apply at "at least this old": with the defaults, a run exactly 26 hours ago already raises the warning.

Collected metrics

• Per job: name, state of the last run, whether the job is disabled, age of the last successful run, age of the newest recovery point, and the number of planned, successful, failed and stopped objects in the last run
• Per repository: name, whether it is attached and consistent, capacity, and free space in GB and percent
• Up to 20 objects with the most failed or stopped runs in the evaluation window
• The number of disabled and never-run jobs, and of past runs with object errors, unexplained skips, long runs and objects without data transfer

Notes

• The check opens a read-only session and verifies that the server honours it before reading anything. Otherwise it aborts without accessing data.
• A disabled or never-run job is reported only when the matching alert option is on. The other checks are skipped for these jobs.
• Schedule-gap detection needs at least four runs in the evaluation window and applies only when the job has no age alert already. For fleets with mixed daily and weekly schedules, set both age thresholds to 0 and rely on schedule-gap detection.

See also: Set up alerts for a check