Notifications and monitoring

A backup job that fails silently is worse than no backup job, because you believe you are protected. This page covers the two things that stop that: notifications, and metrics.

The default is not good enough

Out of the box, Datacenter → Notifications contains two built-in entries:

Name	What it does
mail-to-root	Sends mail to the root account’s address, using the node’s own mail transport
default-matcher	Routes every notification to mail-to-root

On most networks that mail never arrives — nothing is configured to accept it, and Datacenter → Options → Email from address still holds a placeholder sender that mail relays reject.

Assume you are not being told anything until you have proved otherwise

Configure a real target and use Test. Until a test notification actually reaches you, treat every job on the system as unmonitored.

Setting up a notification target

1. Open Datacenter → Notifications.
2. Click Add and choose the target type:

Type	Use it when
SMTP	You have a mail server to send through. The right choice for most sites.
Sendmail	The node’s own mail transport is already configured and working.
Webhook	You want alerts in a chat or on-call system.
Gotify	You use a Gotify push server.

3. Fill in the details — for SMTP that means the server, port, credentials and a sender address your mail system will accept.
4. Give it a clear Target Name and click Add.
5. Select it and click Test.

If the test does not arrive, fix that before doing anything else.

Set the sender address too

Datacenter → Options → Email from address is the address notifications appear to come from. The placeholder it ships with is rejected by most mail systems, and that alone will stop delivery even when the target is otherwise correct.

Deciding what goes where

A matcher decides which notifications reach which target — by severity, by type, by node. The built-in matcher sends everything to the built-in target.

A common arrangement:

Matcher	Sends	To
Failures	Errors and warnings	On-call, via webhook
Everything	All notifications	A team mailbox, via SMTP

That way a failed backup interrupts someone while routine successes stay out of the way.

Notifications from backup jobs

A backup job has its own Notifications tab, with Recipients and a When setting deciding whether mail is sent always or only on failure. Those notifications are delivered through the targets configured here.

Notify on failure only

Daily success mail gets filtered, and then failures get filtered with it. Alert on failure, and check the job list when you want reassurance.

Metrics and long-term history

Every Summary panel has graphs, but they keep a limited window. For longer history, dashboards, or alerting on trends, send metrics to an external system.

1. Open Datacenter → Metric Server.
2. Click Add and choose Graphite, InfluxDB or OpenTelemetry.
3. Enter the server address and port, and any credentials it requires.
4. Give it a name and click Add.

Node and machine metrics are then pushed there continuously.

Which one

Use whichever your organisation already runs. OpenTelemetry is the most portable choice if you are starting fresh, since it is consumed by most modern observability platforms.

What to watch

Signal	Where
Failed tasks	Task panel → Tasks, and node → Task History
Machines with no backup job	Datacenter → Backup → Show: Guests Without Backup Job
Storage filling up	Datacenter → Search, or the storage’s Summary
Thin pool metadata	node → Disks → LVM-Thin
Disk wear and health	node → Disks
Cluster and quorum state	Datacenter → Summary, Datacenter → HA
Replication falling behind	node → Replication, Last Sync column
Licence expiry	node → Subscription

The best five minutes you can spend

Open Datacenter → Backup → Show: Guests Without Backup Job now. It lists every machine no backup job covers. On most systems it is not empty.

If something goes wrong

What you see	What to do
Test notifications do not arrive	Check the sender address under Datacenter → Options, and that your mail system accepts mail from this server.
Some notifications arrive, others do not	A matcher is filtering them. Check its conditions.
Backup failures are not reported	The job’s When setting, or no working target at all. Test the target first.
Metrics do not appear in the external system	Confirm the address and port, and that the node is permitted to reach it — including through the firewall.

Still stuck? Contact VM2Cloud support.