Monitoring & Alerts

Quazzar Space integrates with industry-standard monitoring tools to provide real-time visibility into your infrastructure. This guide covers connecting monitoring platforms, setting up alerts, and using agents for host-level metrics.

Monitoring features are available on Team plans and above.

Supported Monitoring Integrations

You can connect one or more of these tools to your Quazzar Space account. They work independently — use whichever combination fits your existing setup.

Connecting Monitoring Tools

Setup Steps

1. Navigate to Settings > Integrations in your project
2. Select the monitoring tool you want to connect (Prometheus, Grafana, or Zabbix)
3. Enter the connection details:
   • Server URL — The address of your monitoring instance
   • API credentials — API key or username/password for authentication
4. Click Test Connection to verify access. The check reports a list of steps rather than a pass/fail, so you can see exactly where it stopped — a blocked address, a rejected token, and a token that cannot list datasources are three different problems with three different fixes
5. Save the integration

Once connected, Quazzar Space syncs data from your monitoring tools and surfaces relevant metrics alongside your infrastructure inventory.

Prometheus

Connect your Prometheus instance to pull metrics into Quazzar Space. The platform queries your Prometheus server to display metrics for services and servers that match your registered infrastructure.

Grafana

Link your Grafana instance to embed dashboards within Quazzar Space. This provides a unified view without switching between tools.

What the connection check tells you

Testing a Grafana connection runs a sequence of steps and returns all of them, with the ones it never reached marked as skipped:

The last three come from a single /api/datasources call, because Grafana’s own status codes already separate them.

alerting_scope is advisory, and unknown is not ok

alerting_scope never fails the connection. If you use Quazzar Space for metrics and not for alerts, there is no reason to grant a service account an alerting role, and a connection that works for metrics stays healthy either way.

It is asked because silence cannot answer the question. Grafana’s alerting API returns an empty rule list both for a token that can see your whole estate and finds nothing firing, and for a token whose permissions do not reach the folders where your rules live. Nothing in that response is wrong, and no exception is raised — so without asking separately, “nothing is firing” and “we cannot see your alerting” are the same sentence.

The step therefore reports one of three states:

unknown is a third state, not a soft missing and not an ok. An empty rule list from a token whose scope is unknown is reported as unknown, because reporting it as healthy would be claiming an observation we do not hold.

A weaker signal is collected alongside it: the number of folders the token can see. A token can hold the alerting read permission and still be scoped to one folder out of forty, and that count is the only thing that moves when somebody narrows folder access without touching the role — the change that shrinks what is watched while every other check stays green.

Zabbix

Zabbix integration syncs automatically at regular intervals. After connecting:

• Zabbix hosts are mapped to your registered servers
• Zabbix triggers surface as alerts in the Quazzar Space dashboard
• Active problems are displayed alongside your infrastructure status

Creating Alert Rules

Alert rules let you define conditions that trigger notifications when metrics cross a threshold.

Setting Up an Alert

1. Navigate to Monitoring > Alerts in your project or environment
2. Click Create Alert Rule
3. Configure the rule:
   • Name — A descriptive label (e.g., “High CPU on API servers”)
   • Metric — The metric to monitor (CPU usage, memory, disk, response time, etc.)
   • Condition — The operator and threshold (e.g., greater than 85%)
   • Duration — How long the condition must persist before firing (e.g., 5 minutes)
   • Severity — Info, Warning, Critical, or Emergency
4. Select one or more notification channels
5. Save the rule

Severity Levels

Notification Channels

Configure where alerts are delivered. Quazzar Space supports the following channels:

Set up notification channels under Settings > Notifications. You can assign different channels to different alert rules — for example, send warnings to Slack but emergencies to email and webhook.

Service Health Dashboards

Each environment includes a health overview showing:

• Service status — Up, degraded, or down for each registered service
• Recent alerts — Timeline of triggered alerts with severity indicators
• Resource utilization — CPU, memory, and disk usage summaries for monitored servers
• Top consumers — Services and servers using the most resources

This gives your team a single-pane view of infrastructure health without needing to check each monitoring tool individually.

Agent-Based Monitoring

Quazzar Space offers lightweight agents that run on your infrastructure to collect metrics and report health status.

VM Agent

Install the VM agent on your servers to collect:

• Heartbeat — Confirms the server is reachable and responsive
• CPU metrics — Usage percentage, load averages
• Memory metrics — Used, available, cached, and swap
• Disk metrics — Capacity, usage, and I/O statistics
• Service monitoring — HTTP and TCP health checks for services running on the host
• Log collection — Forward logs to Quazzar Space for centralized viewing

Kubernetes Agent

Deploy the Kubernetes agent in your clusters to monitor:

• Cluster health — Node status, pod scheduling, resource pressure
• Pod metrics — CPU and memory per pod
• Service discovery — Automatically detect services running in the cluster
• Event tracking — Kubernetes events surfaced in the Quazzar Space timeline

Agents are configured during installation and report to your Quazzar Space account automatically.

Alert Suppression Windows

During planned maintenance, you can suppress alerts to avoid noise:

1. Navigate to Monitoring > Maintenance Windows
2. Click Schedule Maintenance
3. Select the affected servers, services, or entire environments
4. Set the start and end time
5. Optionally add a description (e.g., “Database upgrade — expected 30 min downtime”)

During the maintenance window, alerts for the selected resources are suppressed. They resume automatically when the window ends.

Plan Requirements

Next Steps

• Infrastructure Management — Organize the resources you are monitoring
• Ansible Automation — Automate remediation when alerts fire
• Cost Optimization — Use utilization data to right-size resources