CNPG Backups Monitoring – Zabbix Template (API-based)

Zabbix Template for CloudNativePG (CNPG) Backups Monitoring on Kubernetes
Automatic discovery of PostgreSQL CNPG clusters, backup monitoring, failure alerts, and backup age thresholds.

This Zabbix template provides full monitoring of CloudNativePG (CNPG) backups through the Kubernetes API, with no custom scripts required on cluster nodes.

🏗 Usage Context

• Kubernetes cluster hosting multiple CNPG databases across several namespaces.
• Zabbix Agent deployed as a DaemonSet for classic node/pod monitoring.
• This template extends monitoring by adding:
  • Automatic discovery of CNPG clusters and backups,
  • Detailed backup checks (status, failures, age),
  • Using only the Kubernetes API (HTTP Agent).

Why use this template?

• Discover & Monitor CloudNativePG (CNPG) PostgreSQL scheduled backups on Kubernetes
• Full Zabbix 7.x integration
• Automatic discovery (LLD) of CNPG clusters
• Backup status (completed / failed / running)
• Backup age monitoring with configurable threshold
• Zero agents or scripts on Kubernetes nodes
• API-only monitoring for maximum reliability

✨ Features

🔍 Low-Level Discovery (LLD)

Backups are discovered via Kubernetes API:

GET /apis/postgresql.cnpg.io/v1/backups

Automatically extracted fields:

• {#NS} → CNPG namespace
• {#CL} → CNPG cluster name

Optional filtering with {$CNPG.NS.FILTER} → in example, to filter namespaces starting with cnpg-*.

📊 Automatically Created Items

For each discovered CNPG cluster:

• cnpg.backup.last_status[{#NS},{#CL}]
  Last known backup status (completed, failed, started, unknown, …)

• cnpg.backup.age[{#NS},{#CL}]
  Age (in seconds) of the last successful backup, or 0 if none has ever succeeded.

🚨 Included Triggers

For each cluster:

• ❌ Backup FAILED — status not completed or started.
• ⚠️ Backup too old — age exceeds {$CNPG.BACKUP.MAXAGE}.
• ⚠️ No successful backup found — no completed backup exists.

✅ Tested With

• Kubernetes: v1.34
• Zabbix Server / Frontend: 7.0.16
• CloudNativePG: ghcr.io/cloudnative-pg/cloudnative-pg:1.27.1

⚙️ Required Kubernetes Permissions (RBAC)

• The default RBAC from zabbix helm chart do allow zabbix to request postgresql.cnpg.io API.
• We need to fix this :

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: zabbix-cnpg-backup-reader
rules:
  - apiGroups: ["postgresql.cnpg.io"]
    resources: ["backups", "scheduledbackups"]
    verbs: ["get", "list", "watch"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: zabbix-cnpg-backup-reader
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: zabbix-cnpg-backup-reader
subjects:
  - kind: ServiceAccount
    name: zabbix-service-account
    namespace: monitoring

To retrieve your zabbix token, use this command:

kubectl -n <NAMESPACE> create token zabbix-service-account

Use it in macro: {$KUBE.API.TOKEN}.

📦 Installation

1. Import the template

Configuration → Templates → Import

Import:

template_cnpg_backups_api.json

2. Assign to Kubernetes API host

• Choose your existing "Kube API hosts" and assign it the template.
• If you d'ont already have a host to monitor your kubernetes cluster using the API, create a new one an set the macros below.
• Macros to configure:

Macro	Description
{$KUBE.API.URL}	Kubernetes API URL
{$KUBE.API.TOKEN}	ServiceAccount token

🔧 Configurable Macros

Macro	Description	Default
{$CNPG.BACKUP.MAXAGE}	Max age of last successful backup	604800
{$CNPG.NS.FILTER}	Filter schedule discovery. Optional	``

📊 Expected Output

• Backup items visible in Latest data
• Alerts appear in Problems for failures/age/no-success

Example of Lastest data

🤝 Contributing

Issues & PRs are welcome.

📜 License

Distributed under the MIT License. See the LICENSE file.